England and Wales High Court (Chancery Division) Decisions

IN THE HIGH COURT OF JUSTICE
CHANCERY DIVISION

		Rolls Building Royal Courts of Justice Fetter Lane, London, EC4AA 1NL
		28/03/2014

Mr Justice Henderson:

Index

Topic	Para
I. Introduction	1-15
II. The factual and procedural background	16-65
(1) Witnesses of fact	16-19
(2) The claimants and the relevant VAT groups	20-23
(3) The underlying facts	24-30
(4) The agreed VAT treatment of agents' commission until 1997	31-35
(5) The disputes and the procedural history from 1997 onwards	36-62
(6) The corporation tax assessments	63-65
III. The decisions in Littlewoods (CA) and Grattan (ECJ)	66-113
(1) Littlewoods (CA)	67-91
(2) Grattan (ECJ)	92-113
IV. Are the Woolwich claims and/or the mistake-based claims, as a matter of English law, and without reference to EU law, excluded by sections 78 and 80 of VATA 1994?	114-115
V. The underlying tax issue: as a matter of EU law, were the VAT payments in question due?	116-149
VI. Estoppel and abuse of process	150-252
(1) Introduction	150-151
(2) Issue estoppel	152-228
(a) General requirements	152-156
(b) The Arnold exception	157-168
(c) The application of issue estoppel to tax cases	169-207
(d) The technical issues	208-228
(i) The same issue	208-213
(ii) Privity	214-228
(3) Contractual estoppel: the 2008 section 85 Agreement	229-242
(4) Abuse of process	243-251
(5) Summary of conclusions	252
VII. The adequate indemnity issue	253-310
(1) Introduction	253-260
(2) British Sugar and Irimie	261-270
(3) Adequate indemnity	271-302
(4) Has the payment of simple interest provided Littlewoods with an adequate indemnity for the loss of use value of the overpaid VAT?	303-310
VIII. Conforming construction or disapplication?	311-341
(1) Introduction	311-316
(2) Conforming construction	317-327
(3) Disapplication	328-341
IX. Change of position and exhaustion of benefits	342-344
X. What is the measure of the claimants' recovery?	345-347
XI. Quantum	348-448
(1) The right approach to quantification of Littlewoods' claims	348-375
(2) The expert evidence of Professor Kay and Dr Richardson	376-407
(a) Government borrowing costs	376
(b) The time value of money	377-384
(c) The actual benefit to the Government from the use of the money	385-407
(3) The views of Vos J in Littlewoods (No. 1)	408-412
(4) Conclusions on the measure of benefit	413-417
(5) Is the Revenue's approach to quantification inconsistent with EU law?	418-425
(6) Should the saving of additional corporation tax be taken into account under English law, and (if so) how should it be computed?	426-448
(a) The position under English law	426-434
(b) Quantification	435-448
XII. Final conclusions	449-450

I. Introduction

1. Over 13 days between 28 October and 13 November 2013 I heard the resumed trial of these two actions, in which the claimants (to which I will often refer collectively as "Littlewoods") seek to recover from the Commissioners for Her Majesty's Revenue and Customs ("HMRC" or "the Revenue") the compounded use value of sums of money which they paid as value added tax ("VAT") between 1973 and 2004. HMRC have subsequently refunded to Littlewoods the principal amounts of those sums, on the footing that they were overpaid, together with simple interest at the statutory rates provided for by section 78 of the Value Added Tax Act 1994 ("VATA 1994"). After allowing for the simple interest received, the total amount claimed as representing the compounded use value of the principal sums repaid, from the dates of the original payments to HMRC until 31 October 2013, comes on the claimants' preferred approach to quantification to no less than £1,212,646,369, or approximately £1.2 billion.
2. The first part of the trial of the two actions ("the First Trial") took place before Vos J (as he then was) on 20 to 23 and 26 April 2010. He handed down his reserved judgment on 19 May 2010: see Littlewoods Retail Ltd and others v Revenue and Customs Commissioners, Littlewoods Ltd and others v Revenue and Customs Commissioners, [2010] EWHC 1071 (Ch), [2010] STC 2072 ("Littlewoods (No. 1)"). In his judgment, Vos J finally determined a number of English law issues relating to liability and the defences available to the Revenue. He also decided that certain questions of EU law relating to liability and remedies needed to be referred for a preliminary ruling to the Court of Justice of the European Union ("the ECJ"), and following a further hearing on 1 and 2 November 2010 he settled the terms of the order for reference in a judgment handed down on 4 November 2010: see [2010] EWHC 2771 (Ch), [2011] STC 171.
3. In Littlewoods (No. 1), Vos J helpfully discussed the main issues of EU law and expressed his preliminary views on them. He also considered some questions of principle relating to the quantification of the claims, and again expressed provisional views on them, in the course of his discussion of the English law defences. I feel sure, however, that he did not intend to rule definitively on any issue of quantum, not least because all such issues had been reserved for the second stage of the trial and he had not yet heard full (if any) argument on them. At a case management conference held on 28 April 2009, Chief Master Winegarten had directed "a trial of all issues of liability in advance of and separately from all issues of quantum (including any related or associated issues of causation)". The First Trial was the trial of those issues, which expressly included HMRC's defence of change of position.
4. Advocate General Trstenjak delivered her Opinion on the questions referred to the ECJ on 12 January 2012, and a Grand Chamber of 13 judges gave its judgment on 19 July 2012: see Case C-591/10, Littlewoods Retail Ltd and Others v Her Majesty's Commissioners for Revenue and Customs, [2012] STC 1714 ("Littlewoods (ECJ)"). Unfortunately, any hopes that this judgment might provide a clear and unambiguous answer to the question of liability which lies at the heart of the present case proved optimistic. Not for the first time, both sides found themselves able to claim, with at least superficial plausibility, that their case had been substantially vindicated. Much argument before me has, of necessity, been devoted to elucidation of the guidance given by the ECJ, and in particular to the question of what the Court meant when it said in paragraph 29 of its judgment:

"In this case, that principle [the EU principle of effectiveness] requires that the national rules referring in particular to the calculation of interest which may be due should not lead to depriving the taxpayer of an adequate indemnity for the loss occasioned through the undue payment of VAT."

5. At a hearing for directions on 17 October 2012, Vos J ordered that there should be "a further hearing for the resumed trial of all outstanding liability issues together with the trial of all quantum issues", and that this trial should take place on the first available date after 1 October 2013. That is indeed the further trial which has now taken place before me, but the position has been complicated by two further developments which I must now mention.
6. The first development is the promotion of Vos J to the Court of Appeal, which was announced in April 2013 and took effect from October 2013. This meant that, unless he could be released from his duties in the Court of Appeal, Vos J would be unable to hear the resumed trial. In the event, he ruled in a judgment which he delivered on 30 April 2013 (see paragraph 10 below) that it was not imperative, although it would be desirable, for him to sit on the resumed hearing: see [2013] EWHC 1310 (Ch) at [21] to [27]. I was subsequently nominated by the Chancellor of the High Court to succeed Vos J as the trial judge, so I found myself in the slightly invidious position of resuming part-heard a trial which had begun before a different judge. I have been greatly assisted in discharging that task by the very full and careful judgments which Vos J has already delivered.
7. The second development is of a more fundamental nature. Both the First Trial and the reference to the ECJ proceeded on the footing, expressly admitted by HMRC in their defences, that the claimants had overpaid the relevant amounts of VAT. As I shall explain in more detail later in this judgment, the liability of HMRC to repay the principal amounts of VAT had been established, first, by the decision of the Court of Appeal in Customs and Excise Commissioners v Littlewoods Organisation Plc [2001] EWCA Civ 1542, [2001] STC 1568 ("Littlewoods (CA)"), and secondly, by HMRC's decision on 20 October 2004 to concede liability in relation to "the 10% element" in appeals which were pending before the Value Added Tax and Duties Tribunal ("the Tribunal"). On top of that, HMRC conceded on the eve of the First Trial that the overpayments of VAT made by Littlewoods had all been made under a mistake of law, and that Littlewoods could not have discovered this mistake until, at the earliest, 26 October 2001 when the decision of the Court of Appeal in Littlewoods (CA) was handed down.
8. Despite this background, however, HMRC applied on 14 March 2013 to amend their defences so as to withdraw their admissions of liability, and their consequential admissions that the payments of VAT had been made under a mistake of law. The justification advanced for this otherwise astonishing volte face was the decision of the ECJ in Case C-310/11, Grattan Plc v Revenue and Customs Commissioners, [2013] STC 502 ("Grattan (ECJ)"), in which judgment had been delivered on 19 December 2012. As I shall explain in due course, the reference to the ECJ in Grattan by the First-tier Tribunal (Tax Chamber) ("the FTT") was made in a factual context similar to that of the present proceedings, but concerned only a question of interpretation of the Second VAT Directive (EC Council Directive 67/228 of 11 April 1967) in relation to periods from 1973 to 1977, before the Sixth VAT Directive (EC Council Directive 388 of 17 May 1977) came into force on 1 January 1978. HMRC nevertheless formed the view, having given careful consideration to the judgment of the Court and to the Opinion of Advocate General Kokott delivered on 13 September 2012, that the reasoning of the Court on this relatively narrow issue led to the conclusion that the sums of VAT in issue in the present proceedings had, after all, been properly due in the first place.
9. Unsurprisingly, the Revenue's application to amend was vigorously opposed by Littlewoods. They argued, among other things, that the decision in Grattan (ECJ) had no bearing on the issues in the present case; that HMRC were estopped from raising their new contentions; and that it would in any event be an abuse of the process of the court if they were permitted to do so, in view of the history of the proceedings and the very late stage at which the application was made. These issues were debated before Vos J at a hearing in late April 2013. By the third day of the hearing, on 30 April, it had become clear to Vos J that, whichever way he decided the application, there was likely to be at least one appeal and/or a further reference to the ECJ, with the almost certain consequence that it would not be possible to keep the October 2013 trial date. He therefore suggested that the parties should agree to defer argument on all points raised by the proposed amendments until the resumed trial. The parties accepted this suggestion, and by paragraphs 1 and 2 of his order dated 3 May 2013 Vos J granted HMRC permission to re-amend their defences and to withdraw the relevant express or implied admissions in their existing amended defences. Paragraph 3 of the order stated that:

"The permissions given in paragraphs 1 and 2 above do not prejudice the claimants' entitlement to raise in their pleadings and at trial any argument relating to issue estoppel or abuse of process which was or might have been raised in opposition to the Commissioners' application issued on 14 March 2013."

10. In the extempore judgment which he delivered on 30 April, Vos J explained his reasons for taking this course as follows:

"4. It seemed to me, as the argument was proceeding, that the far more desirable course was for me to say nothing in a judgment today about the issue estoppel (and, indeed, about the Henderson v Henderson estoppel/abuse of process claims) that are raised by Littlewoods, but to allow HMRC to re-amend its defence in the way that it seeks to do, give Littlewoods permission to amend its reply and, also, possibly, its Particulars of Claim, so that the trial can be a comprehensive trial of all the issues that arise between the parties, including the questions of issue estoppel and abuse of process.

5. This may be a slightly unusual course, but it seems to me that there may well be, if not certainly will be, issues of fact that will arise in relation to the issue estoppel or the extent of the issue estoppel claimed. I took the view, therefore, that it would be far more desirable for all factual and legal issues to be decided at one resumed trial in October 2013 and for any appeals to follow that trial …

6. I, therefore, express no view whatever on the merits of Littlewoods' claims that HMRC should be regarded as issue estopped from raising the claim that the VAT in this case was, in fact, due and payable, notwithstanding their admissions and acceptances in the past that such VAT was not due and payable. I leave all that over to the judge at the trial in October 2013."

11. One of the directions given by Vos J on 3 May 2013 was that the parties should agree a list of issues to be tried. The parties duly exchanged draft lists of issues, and had almost (but not quite) reached agreement on a joint list by the start of the trial. I will use the draft joint list of issues as my framework for most of this judgment, noting where appropriate any differences in the formulations proposed by the parties. For the sake of clarity and completeness, the list helpfully includes issues already decided by Vos J in Littlewoods (No. 1), together with a brief note of what he decided.
12. Before coming on to the issues, however, I must first give a rather fuller account of the factual and procedural background to the case than it was necessary for Vos J to do in Littlewoods (No.1). It is only in the light of this fuller material that the new issues generated by the recent amendments can be properly dealt with. In saying this, I do not wish in any way to detract from the helpful and accurate description of the factual and legal background to the claims already given by Vos J in Littlewoods (No.1) at [10] to [22].
13. I will also take as read the equally helpful "chronology of relevant previous decisions" contained in paragraphs [27] to [43] of Littlewoods (No. 1), and in general I will try to use abbreviations which are either the same as, or similar to, those used by Vos J. As I have explained in a recent case, however, I now prefer to use the description "High Court" rather than "Chancery" to refer to cases decided in the Chancery Division of the High Court. Thus, for example, I will use the description "Chalke (High Court)", rather than Vos J's "Chalke Chancery", to refer to my decision at first instance in Chalke (FJ) Ltd v Revenue and Customs Commissioners [2009] EWHC 952 (Ch), [2009] STC 2027.
14. Vos J's references to the complex FII group litigation ended with the decision of the Court of Appeal in FII (CA): see paragraph [42] of Littlewoods (No.1). Since then, in barest outline, the following major developments have taken place:

i) On 8 November 2010, the Supreme Court granted the parties permission to appeal on a number of issues relating to remedies, while refusing or deferring the grant of permission on other issues, and endorsing the decision of the Court of Appeal to make a further (second) reference to the ECJ on various issues relating to advance corporation tax ("ACT").

ii) On 23 May 2012, the Supreme Court (sitting in a constitution of seven justices headed by Lord Hope of Craighead DPSC) delivered its judgment on the issues for which permission to appeal had been granted: see Test Claimants in the FII Group Litigation v Revenue and Customs Commissioners [2012] UKSC 19, [2012] 2 AC 337 ("FII (SC))". On one important issue of EU law, namely whether EU law protected the test claimants' mistake-based claims, and the related question whether such claims had been validly curtailed by section 320 of the Finance Act 2004, the Court was divided, with the result that a third reference to the ECJ was made.

iii) On 13 November 2012, the Grand Chamber of the ECJ delivered its judgment on the second FII reference. For the purposes of the present proceedings, that judgment is irrelevant and may be ignored.

iv) On 5 September 2013, Advocate General Wathelet delivered his Opinion on the third FII reference, Case C-362/12 ("FII (ECJ) III").

v) By the conclusion of the present hearing on 13 November 2013, the judgment of the ECJ on the third reference was still awaited. If, as then seemed probable, and has since happened, it was delivered before I handed down the present judgment, I gave permission to the parties to file written submissions on any relevant matters which might arise from it.

15. Finally, I should record that on 24 October 2013 I handed down a substantial judgment in two test cases in the CFC and Dividend Group Litigation, which touches on some of the issues in the present case: see The Prudential Assurance Company Ltd and Prudential Holborn Life Ltd v HMRC [2013] EWHC 3249 (Ch), as yet unreported. For the reasons given in paragraph [2] of that judgment, I will refer to it as "Portfolio Dividends (No. 2)". The main parts of the judgment which are relevant to the present case are those dealing with the change of position defence (paragraphs [167] to [193]) and the basis on which interest is payable by the Revenue on the successful claims for unlawfully levied ACT and corporation tax (paragraphs [194 to [247]). In relation to the latter question, I concluded as a matter of English law, following the decision of the majority of the House of Lords in Sempra Metals Ltd v IRC [2007] UKHL 34, [2008] 1 AC 561, ("Sempra"), that the claimants should be awarded compound interest, computed on the conventional basis applicable to government borrowing, on all their successful claims for repayment of unlawfully levied tax. If that conclusion is correct, it was common ground that interest computed on such a basis would satisfy the requirement of effectiveness under EU law, and would provide the claimants with an adequate indemnity within the meaning of paragraph 29 of Littlewoods (ECJ).

II. The factual and procedural background

(1) Witnesses of fact

16. Littlewoods' main witness of fact is Mr Robert Mitchell, who first joined the Littlewoods Group in December 1991 as a VAT controller. Before then, he had been employed by HMRC in various roles since 1971, most recently as the assigned VAT control officer for the Littlewoods Group in relation to its home shopping businesses. In 1994 Mr Mitchell became the Group VAT manager, a post which he continues to hold today. He has given four statements at different stages of the proceedings, of which the second, third and fourth (dated between March and October 2013) are the most material for present purposes. The Revenue did not require Mr Mitchell to attend the trial for cross-examination, so his evidence is unchallenged.
17. The main witness for HMRC on matters relating to the procedural background is Mr Karl Young, a senior lawyer in the VAT and Excise Litigation Department who has conduct of the case on their behalf. As well as dealing with procedural matters, he replied in his third statement dated 12 April 2013 to Mr Mitchell's evidence about the factual background. Mr Young's credibility was not of course in issue, but he was briefly cross-examined by Mr Rabinowitz in order to clarify the nature and basis of his evidence on a few points.
18. Evidence was also given for HMRC by Mr Matthew Richmond and Ms Robina Dyall. Mr Richmond is an economist employed by HMRC who carries out policy analysis for both HMRC and HM Treasury. In that role, he performed various calculations to assist the Revenue's expert on government finance, Dr James Richardson, to prepare his report, and made various other calculations relevant to the cases on quantum advanced by each side. There is no challenge to the accuracy of his calculations, so he was not required to attend for cross-examination.
19. Ms Dyall is a senior civil servant who has been employed by the Revenue for 37 years. Since 2005/6 she has been head of the Administrative Framework Group within the Central Policy department, and as part of that role has had responsibility for the development and maintenance of policy in relation to interest charged on late paid tax and interest paid on repayments of tax. The purpose of her evidence was to give a historical account of the statutory provisions relating to overpaid or underpaid VAT, and to explain the principles which, according to HMRC, underlie the relevant legislation. As with Mr Young, there are no issues about her credibility, but she too was briefly cross-examined by Mr Rabinowitz with a view to probing the real nature of her evidence. It soon became clear that much of it was more in the nature of a departmental policy statement than evidence of facts which Ms Dyall could verify from her personal knowledge and experience.

(2) The claimants and the relevant VAT groups

20. Apart from Littlewoods Limited, which is a holding company, the claimants are all companies which have at different times carried on home shopping businesses. These businesses are typically catalogue-based, and involve sales being made through networks of agents who receive commissions.
21. All of the claimants are currently members of the Littlewoods corporate group, although some are no longer trading. For certain purposes, it is helpful to divide them into two categories. The "Legacy GUS Claimants", namely Reality Group Limited, Shop Direct Group, Kay and Company Limited and Abound Limited, were acquired from the former GUS retail group under a sale and purchase agreement dated 27 May 2003. The remaining claimants, "the Legacy Littlewoods Claimants", have throughout been members of the Littlewoods group.
22. Section 43 of VATA 1994 (re-enacting equivalent provisions in predecessor legislation dating back to section 21 of the Finance Act 1972) makes provision for the grouping of companies together for VAT purposes. Where such a group is formed:

(a) a "representative member" is appointed, and supplies made by or to any member of the group are treated for VAT purposes as being made by or to the representative member;

(b) the representative member is therefore primarily liable to make a VAT return in respect of the collective business of the group, which is treated for VAT purposes as a single economic unit;

(c) the representative member does not pay VAT as agent for the other group members, who are all jointly and severally liable for the tax due from the representative member;

(d) similarly, any claims under section 80 of VATA 1994 for the repayment of VAT paid in respect of the group must normally be made by the group's current representative member; and

(e) where the representative member of a group changes, the replacement company in practice takes over the rights and obligations of its predecessor.

23. Some of the payments in issue were made by claimant companies which did not at the relevant time belong to any VAT group. Most of the payments, however, were made by representative members of VAT groups in respect of liabilities (or supposed liabilities) arising out of the trading activities of the group companies. It is common ground that there are three relevant VAT groups:

(a) The Littlewoods VAT group: Littlewoods Limited has been the representative member of this group since 1973. The other Legacy Littlewoods Claimants belonged to it at different times over the course of the period covered by the claims.

(b) The GUS VAT group: before Littlewoods acquired the Legacy GUS Claimants, they had all (subject to one immaterial point of disagreement) belonged to the GUS VAT group since 1973. This group has had several different representative members over the years. Two of them were acquired by Littlewoods and are claimants in the present proceedings, namely Kay and Company Limited (which was the representative member from 1992 to 1997) and Reality Group Limited (which was the representative member from 1997 to 2003). A further point to note is that when Littlewoods acquired the Legacy GUS Claimants in 2003, the GUS VAT group registration remained with the GUS group, and Experian Finance Plc (then called GUS Plc) became the group's representative member.

(c) The Shop Direct VAT group: when they were acquired by Littlewoods in May 2003, the Legacy GUS Claimants did not immediately join the Littlewoods VAT group. Instead, with the exception of Abound Limited, they joined a newly formed VAT group of which Shop Direct Limited was the representative member. This group was in turn disbanded on 26 June 2004, when Reality Group Limited and Shop Direct Group joined the Littlewoods VAT group. The other two Legacy GUS Claimants, Kay and Company Limited and Abound Limited, did not join, because they had by then ceased trading.

(3) The underlying facts

24. The underlying facts which gave rise to the payments of VAT were briefly summarised by Vos J in Littlewoods (No. 1) at [13] to [15]. A fuller description (set out below) may be found in Littlewoods (CA) at [52] to [56]. It was based on two witness statements of Geoffrey Maitland, a former director of accounting at Littlewoods Retail Ltd. Mr Mitchell says in his third statement that he has read those statements of Mr Maitland "and can confirm that their description of the general operation of the agency arrangements is accurate". It is also common ground that substantially the same system was in operation in the Littlewoods group at all material times since 1973, and that the Legacy GUS Claimants likewise operated a substantially similar model at all material times before they were acquired by Littlewoods in 2003.
25. The relevant paragraphs in Littlewoods (CA) read as follows:

"52. The underlying facts are not in dispute. Littlewoods, through its home shopping division, carries on a retail business selling catalogue goods by mail order through a network of agents. When an agent is recruited to the network he (or, more usually, she) is given an information pack. She is told how to place orders for goods in the catalogue. She is told that she will receive a statement every four weeks which will contain details of all her transactions and will show the minimum amount that she needs to pay. She is told that: "You will receive commission every time you make a payment. You can choose to use your commission in one of two ways – either 12.5% if taken in goods, or 10% if taken in cash".

53. The agent may order goods from the catalogue either for herself or for a third party – usually a friend or colleague. In either case the catalogue price of goods is debited to the agent's account and appears on the statement sent to her. Payments received by Littlewoods from the agent are credited to the agent's account. Those payments may be made by the agent out of her own monies or out of monies which she has collected from the third party customer. An amount of "commission", based on the payments received during the previous four weeks, is credited to a commission account. The accumulated commission is shown as a separate item on the statement. The item takes the form: "Commission available – if taken in goods £[X] or in cash £[Y]." The relationship between X and Y on all the statements that have been put in evidence is such that X is 125% of Y. That reflects the promise in the introductory pack, that: "Commission will be paid on every payment received at a rate of 12.5% if taken in goods, or 10% if taken in cash."

54. The agent "takes" her "commission" by claiming it on one or other of two forms which have been provided to her by Littlewoods. On one of those forms, headed "Commission in Goods 12.5%", she enters particulars of the catalogue goods she wishes to claim, and the total catalogue price. Provided that the total price of those goods (£x) does not exceed the amount (£X) of commission available "if taken in goods", the goods claimed are supplied to her without any monetary payment. The price is debited against her accumulated commission. In such a case the amount of commission available "if taken in goods" is reduced by the total catalogue price of the goods claimed, leaving a balance of £(X-x); and the amount of commission available "if taken in cash" is reduced by a corresponding amount to £(Y-4/5x). If the total price of the goods claimed exceeds the amount of commission available "if taken with goods", the difference, £(x-X), is debited to her account under "Other Items: Charge for Insufficient Commission at Time of Order" and forms part of the new balance owing to Littlewoods for goods supplied. So, in such a case, the goods are supplied to her for a consideration of which the monetary element is part only of the catalogue price.

55. The other form is headed "Commission in Cash 10%". That enables the agent to specify the amount of commission available which she wishes to take "in cash" and offers her the choice between having that amount paid to her by cheque or applied to the credit of her account with Littlewoods. In effect, therefore, she can use her commission to discharge an existing balance on her account; but, if she does so, she is treated as having taken the commission in cash. The amount which she can take in cash (£y) cannot exceed the amount (£Y) of commission available "if taken in cash"; and is debited against her accumulated commission. In such a case the amount of commission available "if taken in cash" is reduced, leaving a balance of £(Y-y); and the amount of commission available "if taken in goods" is reduced by a corresponding amount to £(X-5/4y).

56. As we have said, payments to the credit of the account are made by the agent at four weekly intervals either out of her own monies or out of monies which she has collected from the third party customer. A substantial period of interest free credit is allowed and it is usual for payments to be collected by the agent from third party customers by weekly instalments over a period of, say, 20 weeks. Provided that the third party customer pays all of the instalments due on each item purchased and the agent passes those instalments on to Littlewoods, the agent will be credited, in due course, with commission on the full catalogue price of the third party purchase. But the basis on which the agent is credited with commission is on payments received by Littlewoods from the agent; not on orders placed. This is recognised in those cases where the third party customer fails to keep up instalment payments. In such cases "the agent may invite Littlewoods debt recovery unit to take over the account and collect payments. If this happens the agent will cease to get commission for any further payments received by the debt recovery unit" (see the witness statement of Mr Geoffrey Maitland, formerly financial controller of Littlewoods Home Shopping Division)."

26. Except in relation to periods before 1 January 1978, when article 11C(1) of the Sixth Directive came into force, the disputes between Littlewoods and the Revenue have exclusively concerned commission received by agents in respect of sales to third party customers ("third party purchases" or "TPPs"). It has always been common ground that commission attributable to purchases made by agents on their own account ("agents' own purchases" or "AOPs") should be treated for VAT purposes as a discount from the price of the goods purchased by the agent, rather than as consideration for services supplied by her.
27. Because of the different VAT treatment of commission derived from AOPs and commission derived from TPPs, it was at all material times necessary for the Littlewoods and GUS groups to calculate the percentage of commissions which related to the two types of purchase. The way in which this was done in practice is described by Mr Mitchell in paragraph 30 of his first statement:

"30. In the case of the Littlewoods Group home shopping business, shortly after the inception of VAT in 1973 the group introduced systems which enabled it to identify, at the time orders were placed, which purchases were Agents' Own Purchases and which were Third Party Purchases. In the case of telephone orders the agent was asked when the order was placed whether the goods were for the agent or a customer. When orders were made in writing, the order form used by the agent required them to provide similar information. In this way, the Littlewoods Group analysed each order on an "item-by-item" basis in order to arrive at an Agents' Own Purchases Percentage which could then be applied to the commission earned, to arrive at figures in relation to AOP Commissions and TPP Commissions. In the mid 1990s, the Littlewoods Group adopted a slightly different system, which included returned goods in the analysis, but the principle remained the same."

Mr Mitchell goes on to say that HMRC regularly questioned whether Littlewoods had correctly calculated the AOP percentage. They would sometimes conclude that the percentage had been miscalculated, and would raise a consequential assessment for under-declared VAT.

28. In order to illustrate the kind of system described by Mr Mitchell, Mr Swift referred me to a sample GUS statement book, apparently dating from 1999, which contained a sample form of statement to be completed by the agent when orders were placed. If correctly completed, the statement (which covered a period of 4 weeks) would distinguish clearly between AOPs and TPPs, and between the payments made or received by the agent in respect of each.
29. It is important to note, however, that these records only dealt with the position at the time when the relevant transactions took place. They enabled AOPs to be distinguished from TPPs, and they also enabled calculations to be made of the amounts of commission which agents took in cash or in goods. What they did not reveal, however, at any rate in the majority of cases, was the actual source of commission taken by the agent on any particular occasion. In other words, there was usually no way of telling whether an amount of commission, whether taken in goods or in cash, derived from the agent's own purchases, or third party purchases, or a combination of both. Mr Mitchell explains in his third statement that it is not generally possible, from the account archives maintained by Littlewoods, to differentiate between orders placed by an agent for a third party and orders placed by an agent for herself, or between payments made in respect of the two types of order: see paragraphs 21 and 22.
30. Mr Mitchell goes on, in paragraphs 38 and following of his third statement, to describe the nature of an agent's commission account, and he draws the following unchallenged conclusions:

"40. It is clear from the Account Records that commission is generated from, and calculated by reference to, qualifying payments made by the agent to the mail-order companies. An agent does not become entitled to commission by placing an order …

41. Moreover, it is clear from the account records that the commission is generated without regard to (i) whether the payments were made in respect of the agent's own purchases or those made on behalf of third parties or (ii) whether the payments were made from the agent's own money or that collected from third parties. Similarly, the account does not reconcile any payment made by an agent to any particular order. Furthermore, the agents' rights to commissions were not dependent on whether they had arranged any sales to third party customers at all.

42. It is also clear that the commission account is independent of the trading account. The commission account balance is not automatically set off against the outstanding trading account balance. Instead, the agent can elect when and how to use the commission."

(4) The agreed VAT treatment of agents' commission until 1997

31. From the inception of VAT in the United Kingdom on 1 January 1973 until May 1997, there was no relevant disagreement between Littlewoods (or the GUS group) and the Revenue about the correct VAT treatment of commission earned by agents.
32. As I have already explained, it was agreed that commission earned in respect of agents' own purchases should be treated as a discount from the price of goods purchased by the agent. All that an agent had to do in order to earn commission on goods which she had purchased for her own account was to make a payment to the mail order company of, or on account of, the purchase price. No third party was involved, and the agent was herself the final consumer for VAT purposes of the goods which she had purchased. It is therefore easy to see why the so-called commission derived from the payments made for her own purchases ("AOP commission"), where it was taken in cash, was characterised as a retrospective discount, or rebate, from the catalogue price of the AOPs which she had previously made. Where AOP commission was taken in goods, it was treated as an immediate discount at the time of supply from the catalogue price of the "secondary" goods ordered by the agent, on the footing that the consideration received by the company was reduced by the amount of the commission.
33. Commission attributable to third party purchases ("TPP commission"), however, was treated differently, with a further distinction again depending on whether the commission was taken in goods or in cash. Where the TPP commission was taken in cash, the whole of it (at the rate of 10%) was treated by the Revenue as consideration for services provided by the agent to the mail order company. Again, this is in principle easy to understand. In the case of third party purchases, unlike purchases for her own account, the agent had to find the purchaser, arrange the sale, and collect payment for transmission to the vendor. In respect of such purchases, she was in effect acting as a commission agent, and the commission was her remuneration for so acting. Thus characterised, the TPP commission was the consideration for a supply of services by the agent to the company. It could not simultaneously be treated as reducing the consideration for a supply of goods by the company to the agent.
34. Where the TPP commission was taken by the agent in goods, a distinction was drawn between the basic element of 10% and the additional element of 2.5%. The 10% element was still treated as consideration for services provided by the agent to the vendor, but the additional 2.5% element (amounting to 20% of the total commission) was treated as a discount from the price of the secondary goods purchased by the agent. The reason for the distinction, presumably, was that the 10% element of the commission was thought to represent consideration for the agent's services in exactly the same way as when the commission was taken in cash, whereas she had done nothing further in order to earn the additional 2.5%. It therefore seemed appropriate to regard the 2.5% as a discount from, or reduction in, the catalogue purchase price of the secondary goods which the agent ordered when she decided to take her commission in goods.
35. In accordance with the treatment which I have described, the Littlewoods and GUS companies always accounted for VAT on the full purchase price of goods sold through the agent to third party purchasers. However, in relation to secondary goods sold to the agent when the commission thus taken was itself attributable to third party purchases, VAT was accounted for on the basis of the catalogue price of those goods reduced by the additional 2.5% element referred to in paragraph 34 above.

(5) The disputes and the procedural history from 1997 onwards

36. On 16 May 1997, HMRC wrote to Littlewoods Ltd (then called The Littlewoods Organisation) saying they were no longer satisfied that the treatment of TPP commission taken in goods was correct, and that for the future Littlewoods should account for such supplies on the basis of the catalogue selling price, i.e. without a 2.5% discount. In a similar letter to GUS Home Shopping Ltd dated 13 May 1997, HMRC explained their change of position as follows:

"The additional 2.5% is not a freely given discount, the agent has supplied her selling services in order to be entitled to the option of taking her 12.5% commission in this way. If the £2.50 is not considered to be money received in payment for the goods, then it must be non-monetary consideration.

…

Suppose the original supplies of £200 are made. Assuming that half of these are bought by customers and half by the agent herself, this would give an AOP percentage of 50%. Tax would have been accounted for on the consideration of £190. Commission in the commission account would be £20. If the agent chose to take the commission/discount earned as goods, the value of commission is increased to £25. The amount of discount earned would increase by 2.5%, hence tax would be due on £187.50 plus the goods purchased with commission, £25.00.

The value upon which tax has been accounted for on GFC [goods for commission, or secondary goods] it too low as it fails to include the non-monetary element of the consideration received by the Company from the agent for the goods."

37. In other words, the argument (as I understand it) was that where TPP commission was applied in the purchase of secondary goods, the whole of the discount fell to be treated as consideration for the agent's services, and not just the 10% element. In the letter to Littlewoods Ltd, this revised treatment was said to be consistent with the recent decision of the Court of Appeal in Rosgill Group Ltd v Customs and Excise Commissioners [1997] STC 811. In that case, the Court had held that the non-monetary element of the total consideration for the supply of a blouse to a hostess entitled to a discount on her own purchases of clothing under the so-called "party-plan" method was the value which the parties put on her services in the events which happened, namely her choice to buy the blouse at a discount rather than be paid a smaller amount in cash: see the discussion of the case in Littlewoods (CA) at paragraphs [25] to [28].
38. Littlewoods did not agree with HMRC's revised stance, and continued to account for VAT on the sale of secondary goods on the basis that the 2.5% element represented a discount. HMRC then issued assessments in relation to the tax they considered to have been under-declared, and an appeal against the assessments came before the Tribunal. The Tribunal released its decision on 19 October 1999, allowing Littlewoods' appeal.
39. HMRC then appealed to the High Court, where Lightman J reversed the Tribunal's decision: see Customs and Excise Commissioners v Littlewoods Organisation Plc [2000] STC 588. Lightman J held:

(a) that the agent's services constituted part of the consideration for the supply of the secondary goods, on the basis that the necessary direct link existed between the provision of the services and the exercise by the agent of her option to take the relevant commission in goods; and

(b) following the decision of the Court of Appeal in Rosgill, that the 2.5% element formed part of the consideration for the agent's services.

40. On Littlewoods' further appeal to the Court of Appeal, the decision of the Tribunal was reinstated. The Court of Appeal delivered its judgment in Littlewoods (CA) on 26 October 2001, following the combined hearing over four days in July 2001 of four appeals which raised (or in one case were thought to raise) related questions about the determination for VAT purposes of the monetary equivalent of the consideration for a supply of goods, where the supply was made in whole or in part for a non-monetary consideration. The judgment of the Court (Aldous, Chadwick and Sedley LJJ) was delivered by Chadwick LJ. The judgment was a conspicuously thorough one, which started from first principles and then reviewed the relevant case law of the ECJ, as well as relevant domestic decisions of the Court of Appeal, before turning to the individual circumstances of the four appeals.
41. I will need to consider the Court of Appeal's reasoning in relation to Littlewoods' appeal in much greater detail later in this judgment. The important point to note at this stage is that, although only the 2.5% element of TPP commission taken in goods was in issue on the appeal, the reasoning of the Court seemed to entail that the 10% element of TPP commission should also be treated as a discount, and furthermore that this was so whether the commission was taken in goods or in cash. In particular, the Court appeared to endorse the following key principles:

(a) First, the 2.5% element could not be separated from the 10% element where TPP commission was taken in goods. Either the entire 12.5% was a discount, or none of it was: see [68] and [71].

(b) Secondly, the entire 12.5% had to be treated as a discount, and no part of it could constitute consideration for a supply of services, because there was no direct link between any services supplied by the agent and the supply to her of secondary goods for less than the full catalogue price: see [70] to [75].

(c) Thirdly, while not deciding the point, the Court said at [67] that it found it difficult to see why TPP commission taken in cash should not likewise be treated as a discount.

42. Despite suffering this comprehensive defeat, the Revenue decided not to seek leave to appeal to the House of Lords. They had been represented at the hearing by very experienced leading counsel (Mr Nigel Pleming QC), and the decision must have been taken after careful consideration. Indeed, the Revenue confirmed as much in a letter sent to Mr Mitchell on 18 February 2002, which said that the decision not to appeal had been taken "after due consideration". It appears that a strategic decision was instead made to limit the damage by attempting to confine the decision of the Court of Appeal to the treatment of the 2.5% element of TPP consideration taken in goods which had actually been in issue.
43. On 5 February 2002, Littlewoods made a claim under section 80 of VATA 1994 for repayment of the 10% element of all TPP commission, whether it had been taken in cash or in goods. The claim was, however, limited to the three years from January 1999 to December 2001, in accordance with the three year limitation period in section 80, the lawfulness of which under EU law was currently under challenge in the Marks and Spencer litigation.
44. In the letter of claim, Mr Mitchell summarised Littlewoods' position as follows:

"Stemming from the comments of the Court, and following further legal advice, we [now] believe that for VAT purposes no distinction should be drawn between commission earned on agents' own purchases and that earned on third party purchases. Moreover, it is our view that, as there is no direct link between commission paid out and the service provided, all the commission paid out in cash or taken as further goods should be treated as a discount under article 11C(1)."

45. HMRC responded to this letter on 18 February 2002, rejecting the claim. The senior officer who wrote the letter, Stewart Bain, adopted the policy of damage limitation to which I have already referred. He argued that there was no evidence in the Court of Appeal's judgment that it had found the necessary link to be absent between the services provided by the agents in making sales to third parties and the 10% element of the commission attributable to such sales. He said it was the considered view of the Revenue, having decided not to appeal, "that the decision of the court should not be extended to circumstances which were beyond the boundaries of the original appeal".
46. On 19 March 2002, Littlewoods Ltd appealed to the Tribunal against the decision rejecting the claim. The resulting proceedings ("the 10% Commission Appeal") continued for some two and a half years, with HMRC maintaining their position both in correspondence and in their statement of case, until on 20 October 2004, about two weeks before the scheduled hearing of the appeal, they decided to capitulate. In a letter of that date, the Solicitor's Office of HM Customs and Excise wrote as follows:

"Having considered this matter very carefully, in consultation with leading counsel, the Commissioners have decided to withdraw the disputed decision presently under appeal.

May I apologise to you and your client for the time it has taken to reach this conclusion but I am sure you will appreciate the Commissioners are under a duty to take particular care and give detailed consideration to issues that have large revenue implications."

47. Before reaching this last minute decision to climb down, the Revenue had undertaken detailed factual investigations (described by Mr Mitchell in paragraphs 48 and 49 of his second statement) with a view to gaining a better understanding of how Littlewoods' commission arrangements with their agents actually operated. In particular, the Revenue's representatives wished to investigate whether a factual distinction could be drawn between the 10% and 2.5% elements of TPP commission, because no evidence specifically directed to the 10% element had been before the Court of Appeal.
48. From the Revenue's point of view, these investigations were not a success. According to Mr Young, "they did not open up any significant distinction" between the facts which had been before the Court of Appeal in Littlewoods (CA) and the facts in the 10% Commission Appeal. Accordingly, there was perceived to be no realistic prospect of establishing a direct link between the 10% element and services provided by the agents, any more than there had been in relation to the 2.5% element.
49. The amount of VAT reclaimed in the 10% Commission Appeal itself was approximately £4.4 million. By the time when the appeal was conceded in October 2004, the following further claims had been submitted in relation to the 10% element:

(a) On 25 June 2002, Littlewoods Ltd made a claim covering the years 1973 to 1998, relying on the opinion of the Advocate General in Marks and Spencer. HMRC rejected the claim, and on 19 September 2002 Littlewoods Ltd appealed to the Tribunal, estimating the amount of tax in dispute to be £50 million. This appeal was then adjourned pending the outcome of the 10% Commission Appeal.

(b) On 1 June 2004, Littlewoods Ltd made a further claim to recover about £2.8 million for periods from 2002 to 2004. On 8 June 2004, HMRC rejected the claim and it was then held in abeyance pending determination of the 10% Commission Appeal.

(c) On 24 June 2003, Experian Finance Plc (then called GUS Plc) made a claim for about £122 million as the (then) representative member of the Argos VAT group for periods from 1973 to 2002. The claim was made at the instance, and for the benefit, of Littlewoods, as one of the terms of the acquisition of the Legacy GUS Claimants by Littlewoods in May 2003 had been that Littlewoods would get the benefit of any VAT recovery claims associated with their past trading. One of the first things which Mr Mitchell had done, after taking over responsibility for their VAT affairs, had been to check whether section 80 claims had been lodged in respect of the 10% element and to ensure that this was now done. The claim was apparently neither accepted nor rejected by HMRC, but left in abeyance pending the 10% Commission Appeal.

(d) On 27 October 2004, a supplemental claim was made covering periods from April 2002 to September 2004.

50. It is clear, therefore, that when HMRC conceded the 10% Commission Appeal they were well aware of the possible financial implications, even though the invalidity of the three year limitation period in section 80(4) of VATA 1994 was not finally established until after the decision of the House of Lords in Fleming v Customs and Excise Commissioners [2008] UKHL 2, [2008] 1 WLR 195. Furthermore, documents disclosed by the Revenue show, as one would expect, that the wider financial implications of the 10% Commission Appeal were internally discussed and understood. For example, in a policy paper submitted on 12 August 2003 by two senior officers involved in the 10% Commission Appeal, a discussion of evidential issues and of the need to establish a "direct link" concluded as follows:

"For background information, the revenue at stake in this issue across the four large mail order companies in the UK is [figure redacted]. Claims have been received from these companies covering the past three years (and back to 1973 in some cases) and of course this is tax that is at risk going forward."

51. The terms of settlement of the 10% Commission Appeal were set out in a written agreement expressed to be governed by section 85 of VATA 1994, subsection (1) of which (as then in force) provided that:

"Subject to the provisions of this section, where a person gives notice of appeal under section 83 and, before the appeal is determined by a tribunal, the Commissioners and the Appellant come to an agreement (whether in writing or otherwise) under the terms of which the decision under appeal is to be treated –

(a) as upheld without variation, or

(b) as varied in a particular manner, or

(c) as discharged or cancelled,

the like consequences shall ensue for all purposes as would have ensued if, at the time when the agreement was come to, a tribunal had determined the appeal in accordance with the terms of the agreement (including any terms as to costs)."

52. So far as material, the agreement ("the 2004 section 85 Agreement") provided as follows:

"WHEREAS [HMRC] have rejected the Appellant's claim, pursuant to section 80 of the VAT Act 1994, for the recovery of output tax overpaid in the sum of £4,443,162.12, contained in a letter dated 5 February 2002 ("the Claim")

NOW in consideration of the terms set out herein IT IS HEREBY AGREED between the parties as follows:

1. [HMRC] will repay the Appellant's Claim, subject to the receipt of satisfactory evidence to support the quantum and method of calculation of the Claim.

2. … "

53. Since the settlement of the 10% Commission Appeal, HMRC have now repaid to Littlewoods principal amounts totalling £204,774,763. They have also paid simple interest on these amounts, at the statutory rate prescribed by section 78 of VATA 1994, of £268,159,135.
54. HMRC have withheld payment of parts of the claims totalling £36,226,663 on account of three issues, one of which Littlewoods has accepted, but two of which remain in contention and have yet to be determined in other proceedings. The point which Littlewoods no longer disputes is referred to in the statements of case as "the Freemans point", and I need say no more about it. The other two issues are:

(a) the prospective capping point; and

(b) the GMAC point.

55. The prospective capping point concerns VAT paid in periods ending after 4 December 1996 and up to 2 January 1999 (in the case of the Littlewoods VAT group) and 31 March 2000 (in the case of the GUS VAT group). HMRC argue that section 80(4) of VATA 1994 is still effective to exclude claims for the repayment of those amounts. The point remains in issue, but Littlewoods' appeal on it has been stayed behind other proceedings.
56. The GMAC point concerns all periods before 1978, when the Second VAT Directive was in force. HMRC rejected claims to recover VAT paid in respect of those periods on the basis that, before article 11C(1) of the Sixth Directive came into force on 1 January 1978, there was no authority under EU law for taxable persons to adjust output tax when the consideration for a supply was subsequently reduced. In support of this contention, HMRC rely on a decision of the Tribunal in General Motors Acceptance Corporation (UK) Plc v HMRC, ("GMAC"), VAT 19989 (14 February 2007). HMRC originally made deductions on this basis in respect of commission taken both in goods and in cash. As Mr Mitchell explains in his fourth statement, appeals against the relevant decisions were made by the appropriate claimants in October 2007. Following representations on behalf of Littlewoods, on 27 June 2008 HMRC accepted that the logic of their argument did not cover cases in which commission was taken in goods. They therefore agreed to repay the amounts of VAT attributable to commission so taken, and this agreement was recorded in a further section 85 agreement dated 23 September 2008. Pursuant to this agreement ("the 2008 section 85 Agreement"), HMRC made further repayments of approximately £4.5 million principal and £12.8 million interest. These repayments are included in the total figures given in paragraph 53 above. This left in dispute amounts of VAT paid before 1978 in respect of commission taken in cash ("the remaining GMAC point"). Littlewoods' appeals in relation to that issue have been stayed behind other proceedings brought by Grattan Plc in which the same point arises (and in which the ECJ has now given its ruling in Grattan (ECJ), discussed further below).
57. The present proceedings were begun by claim forms issued on 13 March and 5 October 2007. Nothing turns on the existence of two separate actions, which appears to be a historical accident mainly caused by the way in which the section 80 repayment claims were originally made to the Revenue. The first action relates only to payments of VAT made in and after 1999 and 2000, while the second action mainly relates to earlier payments dating back to 1973 (although it also includes claims made by Littlewoods Ltd in respect of payments made between 1999 and 2004).
58. As I have already explained, most of the claims relate to periods when the companies in question were organised into VAT groups, although some of the claims are made by companies which were separately registered for VAT when the relevant payments were made. Where payments were made by a representative member in respect of other group companies, there might have been a question about the identity of the correct claimant in a common law restitution claim. Accordingly, all such claims have been pleaded in the alternative on behalf of:

(a) the companies whose trading activities gave rise to the supposed VAT liabilities;

(b) the representative members which made the payments, or their assignee; and

(c) the current representative member of the relevant group, or its assignee.

It is now common ground, however, for the purposes of the present claims only, that the correct claimants are the representative member companies which actually paid the sums in question. The parties have therefore been able to agree a list of the correct claimants. They include Shop Direct Group, which has taken assignments from three of the former representative member companies of the GUS VAT group which are not claimants in their own right: see Littlewoods (No. 1) at [20].

59. The claimants advance both mistake-based and Woolwich restitutionary claims to recover the full time value of the allegedly overpaid VAT. Much of the disclosure and witness evidence for the First Trial related to the question whether the relevant payments had been made by mistake. On 19 April 2010, the day before the trial began, HMRC belatedly accepted that the payments of VAT had all been made under a mistake of law, and that Littlewoods could not have discovered the mistake until, at the earliest, the decision in Littlewoods (CA) on 26 October 2001. As a result of these concessions, it was common ground at the First Trial that the necessary ingredients of the mistake-based claims were all present, and that none of the claims was time-barred, although HMRC's primary argument was that the claims were all excluded as a matter of English law by section 80 of VATA 1994.
60. It was also common ground that, subject to the same primary argument, the claimants' alternative Woolwich claims were bound to succeed, although only in relation to payments made within six years of the dates of issue of the claim forms.
61. If the claimants are entitled to assert their mistake-based claims, which reach back to 1973, they do not need to rely on the alternative Woolwich claims with their much shorter limitation period. The claimants will, however, be confined to their Woolwich claims if:

(a) as Vos J has now held, section 80 excludes the claimants' common law claims as a matter of English law;

(b) it is necessary for the court to disapply or modify that exclusion so as to give effect to the claimants' directly effective EU law rights; and

(c) the court were to agree with the provisional view expressed by Vos J to the effect that, in such circumstances, the claimants would only be able to assert their Woolwich claims (see Littlewoods (No. 1) at [77] to [92]).

The second and third of these questions are among the issues which I will have to decide.

62. In the introductory section of this judgment, I have already dealt with the procedural developments since the date of the First Trial which have led the Revenue to seek to re-open the underlying issue of liability to pay the VAT in the first place. If those contentions were to succeed, the result would be to defeat the present claims in their entirety, or at least those claims in respect of which the Revenue were permitted to argue the point. HMRC have made it clear, however, that they will not in any circumstances seek to recover the principal sums and simple interest already paid to the claimants. In other words, they wish to rely on the liability argument only as a defence to the present claims.

(6) The corporation tax assessments

63. When a repayment of VAT and statutory interest has been made, HMRC contend that both elements of the repayment (the principal sum and the statutory interest) are chargeable to corporation tax in the hands of the recipient, either as a receipt of its trade (if it is still trading) or as a post-cessation receipt (if it has ceased trading). In broad terms, the argument is that the repayments arise from the carrying on of a trade by the company, and they fill a gap, or hole, in the company's income which was caused by the original overpayments of VAT. Various issues arise, including questions about the taxable source of the payments, about the chargeability of the interest under the legislation relating to loan relationships or otherwise, and about the relevance of the VAT group structures through which the tax was paid and repaid.
64. In order to test these issues, HMRC made amendments to the corporation tax self-assessments of two of the present claimants, Shop Direct Group and Shop Direct Home Shopping Ltd, from which those companies appealed to the FTT. Their appeals (together with those of two other companies raising similar issues) were dismissed by the FTT (Judge Berner and Miss O'Neill) in a decision released on 14 February 2012: see Shop Direct Group v HMRC [2012] UKFTT 128 (TC), [2012] SFTD 723.
65. The taxpayer companies then appealed to the Upper Tribunal, but on 19 April 2013 their appeals were dismissed by Asplin J: see Shop Direct Group and others v Revenue and Customs Commissioners [2013] UKUT 189 (TCC), [2013] STC 1709. Shop Direct Group, alone of the appellants, sought and was granted permission to appeal to the Court of Appeal, but only in relation to one of the repayments of VAT and its associated interest payment. This appeal was heard in February 2014 and dismissed on 11 March 2014: see [2014] EWCA Civ 255. The litigation has thus determined in HMRC's favour all of the points of principle relating to the taxability of the sums of principal and interest already paid.

III. The decisions in Littlewoods (CA) and Grattan (ECJ)

66. In view of their central importance to the Revenue's new argument on liability, I will in this section of my judgment review the decisions of (a) the Court of Appeal in Littlewoods (CA), and (b) the ECJ in Grattan (ECJ).

(1) Littlewoods (CA)

67. After setting out the relevant legislative provisions of the Second and Sixth VAT Directives, the Court embarked on a detailed historical and thematic review of the general principles underlying the appeals. They began with the fundamental principles established by the ECJ in the "Dutch Potato" case, Staatssecretaris van Financiën v Coöperatieve Aardappelenbewaarplaats GA (Case 154/80), [1981] ECR 445. The Court recorded at [14] that three principles find expression in the judgment of the ECJ in this case:

"(i) there must be a direct link between the supply of goods or services and the consideration which is said to have been received for that supply;

(ii) the consideration must be capable of being expressed in money or a monetary equivalent; and

(iii) the basis of the assessment is the consideration actually received."

68. In relation to the third principle, the Court of Appeal explained in the same paragraph that:

"The "subjective" value must be ascertained by reference to the consideration actually received for the goods or services actually supplied. The enquiry excludes any valuation which is independent of the actual transaction; that is to say, any valuation based on criteria which are not those adopted by the parties themselves."

69. The Court then reviewed the affirmation and application of those three principles by the ECJ in two subsequent cases, Naturally Yours Cosmetics Ltd v Customs and Excise Commissioners (Case 230/87), [1988] ECR 6365, [1988] STC 879 ("Naturally Yours") and Empire Stores Ltd v Customs and Excise Commissioners (Case C-33/93), [1994] ECR I-2329, [1994] STC 623 ("Empire Stores"). In the light of that review, the Court at [23] refined the third of the "Dutch Potato" principles as follows:

"In the light of the later decisions it can be seen that, in cases where the consideration for the supply of goods does not consist (or does not wholly consist) of money, there are at least two factual situations in which it may be necessary to ascertain the monetary equivalent of the non-monetary element of that consideration. The first is where the parties have, as between themselves, put a value on the non-monetary element. In those cases – of which the Naturally Yours case is an example – the monetary equivalent is to be taken at the value adopted by the parties. The second factual situation is where the parties have not dealt with each other on the basis of consensus as to the monetary equivalent of the non-monetary element. In those cases – of which the Empire Stores case is an example – the monetary equivalent has to be ascertained in some other way … In such cases, the monetary equivalent of the services supplied is the amount that the recipient of the services is prepared to pay for the purpose of obtaining those services."

70. The Court of Appeal then discussed:

(a) its own decisions, applying the principles developed by the ECJ, first in the Rosgill case, and then in Customs and Excise Commissioners v Westmorland Motorway Services Ltd [1998] STC 431 (see [24] to [31]);

(b) two cases in the ECJ dealing with problems where goods were supplied against redeemable coupons previously issued by the supplier (in paragraphs [32] to [40]); and

(c) further questions which arise where the customer is allowed a discount or rebate, not at the time of supply, but at a time after the relevant supply has taken place (in paragraphs [41] to [49]).

71. Cases of the third type engage the provisions of article 11(C)(1) of the Sixth Directive, which provides that:

"In the case of cancellation, refusal or total or partial non-payment, or where the price is reduced after the supply takes place, the taxable amount shall be reduced accordingly under conditions which shall be determined by the Member States."

Under this last heading, the Court discussed the important decisions of the ECJ in Elida Gibbs Ltd v Customs and Excise Commissioners (Case C-317/94), [1996] ECR I-5339, [1997] QB 499 ("Elida Gibbs") and Freemans Plc v Customs and Excise Commissioners (Case C-86/99), [2001] ECR I-4167, [2001] 1 WLR 1713 ("Freemans").

72. As the Court of Appeal pointed out in its discussion of Elida Gibbs, the ECJ in that case "took the opportunity to restate the basic principles of the VAT system" ([44]). Thus, in paragraph 19 of its judgment the ECJ emphasised that:

"The basic principle of the VAT system is that it is intended to tax only the final consumer. Consequently the taxable amount serving as a basis for the VAT to be collected by the tax authorities cannot exceed the consideration actually paid by the final consumer which is the basis for calculating the VAT ultimately borne by him."

73. Another fundamental principle was that article 11(C)(1) of the Sixth Directive "had to be interpreted so as to give effect to the principle of neutrality" ([45]). It followed from this principle that the taxable amount of goods supplied to a final consumer had to be reduced by the amount of a voucher or coupon given to him by the first company in the supply chain, even though there was no contractual relationship between them. As the ECJ explained in paragraph 31 of its judgment:

"… in order to ensure observance of the principle of neutrality, account should be taken, when calculating the taxable amount for VAT, of situations where a taxable person who, having no contractual relationship with the final consumer but being the first link in a chain of transactions which ends with the final consumer, grants the consumer a reduction through retailers or by direct repayment of the value of the coupons. Otherwise, the tax authorities would receive by way of VAT a sum greater than that actually paid by the final consumer, at the expense of the taxable person."

74. The underlying facts in Freemans were very similar to those of the present case, but the issue referred to the ECJ by the Tribunal was purely a timing one about when the discount on agents' own purchases should be taken into account. As the Court of Appeal said at [47]:

"The question referred to the Court of Justice for a preliminary ruling was whether the taxable amount in respect of goods supplied to the agent for the agent's own use was (1) the full catalogue price of the goods supplied less the AOP discount on that price or (2) the full catalogue price with a reduction as and when the AOP discount was credited to the agent or (3) the full catalogue price with a reduction as and when the AOP discount was withdrawn or used by the agent. It should be noted that the question referred was restricted to the treatment of the AOP discount – that is to say, the court was not asked to consider, separately or at all, the proper treatment of commission in respect of purchases made for other customers."

The answer given by the ECJ was that the third possible solution was the correct one, with the result that the taxable amount of the supply was reduced only when the AOP discount was withdrawn or used by the agent.

75. Having thus set the scene, the Court of Appeal then dealt with Littlewoods' appeal. I have already set out the Court's statement of the underlying facts in [52] to [56]: see paragraph 25 above. The Court next identified the specific issue which it had to consider. It recorded at [57] that there was no dispute in relation to the VAT treatment of third party purchases. Nor was there any dispute about the treatment of agents' own purchases, although following the decision of the ECJ in Freemans it was now common ground that, where the relevant commission was taken in cash, it was only at that point that the taxable amount of the goods supplied to the agent could be reduced (Littlewoods had previously sought to treat such goods as supplied at a discount of 10%, and had initially accounted for VAT on only 90% of the catalogue price).
76. The Court of Appeal continued, at [58]:

"The dispute is as to the treatment, for the purposes of VAT, of the supply of goods to the agent herself in those cases where the commission which she has earned is "taken in goods" – that is to say, where goods are supplied to her in response to a claim on a "Commission in Goods 12½%" form. In such cases the Commissioners have accepted that the taxable amount of those goods should be the catalogue price reduced by an amount equal to the commission claimed, but at the "taken in cash" rate of 10%. It seems that this treatment is thought appropriate on the basis that it would be wrong to refuse a reduction equal to that which would be allowed if the commission were in fact "taken in cash". The issue is whether a further reduction of 2.5% should be allowed; or, more precisely, whether that further reduction should be allowed in so far as the commission "taken in goods" is commission attributable to payments received in respect of sales to third parties. Littlewoods contends that it should be treated as an additional discount. The Commissioners argue that it should be treated as the monetary equivalent of non-monetary consideration provided to Littlewoods, as supplier, by the agent – such non-monetary consideration being the agent's services in introducing, and procuring sales to, third party customers."

77. Although this paragraph was intended to identify the actual issue on the appeal, the second and third sentences (which I have italicised) are, with the greatest respect, erroneous. As I have already explained, HMRC had never accepted that there should be a 10% discount from the purchase price of secondary goods ordered by the agent, nor was any reduction allowed if the commission was taken in cash. In each case, the 10% element was regarded by both HMRC and Littlewoods as non-monetary consideration for the agent's services in relation to the third party purchases from which the commission derived, and as forming part of the consideration for the supply of the secondary goods. Furthermore, this error clearly had an influence on the formulation by the Court of the issue which it actually had to decide, because in the fourth sentence of [58] that issue was said to be whether a "further" reduction of 2.5% should be allowed, and in the next sentence Littlewoods' contention was recorded as being "that it should be treated as an additional discount" (my emphasis). In reality, however, the issue between the parties was not whether the discount should be 12.5% instead of 10%, but rather whether (a) the entire 12.5% should be regarded as consideration for the agent's services, or (b) the 10% element should be so regarded, but the additional 2.5% should be treated as a discount.
78. After referring to the decision of the Tribunal and the judgment of Lightman J, the Court continued:

"63. Littlewoods appeals from the order made by Lightman J with the permission of this court (Mummery LJ) granted on 8 August 2000. The grounds of appeal … are that the judge –

"… erred in deciding that when an Agent purchased further goods using some or all of the commission she had earned, the value attributed by Littlewoods to the non-monetary consideration provided by the Agent in the form of her services in earning that commission was enhanced from 10% to 12.5%"

In the skeleton argument lodged on its behalf (before the decision of the Court of Justice in the Freemans case) it is said that the judge ought to have held that the enhanced value of the commission (the additional 2.5%) was properly to be treated as a price reduction or discount within art 11A(3) of the Sixth Directive."

This paragraph, by contrast, accurately reflects the true area of dispute between the parties. Littlewoods' contention was that Lightman J had erred in holding that the services element of the commission was enhanced from 10% to 12.5%, and the skeleton argument in support of the appeal said only that the additional 2.5% element was properly to be treated as a price reduction or discount.

79. The Court then returned at [64] to the issue on the appeal, but reformulated it in wider terms which required an examination of the whole amount of the commission, not just the 2.5% element:

"64. The issue on this appeal is how the taxable amount in respect of the supply of goods to the agent herself, in response to an order made on a "Commission in Goods - 12½% Form", is to be ascertained. The relevant question is whether the difference between the catalogue price for the goods ordered and the monetary amount (if any) paid by the agent, or debited to her account as "Charge for Insufficient Commission at Time of Order", is to be treated as a price discount allowed to the customer and accounted for at the time of the supply – within art 11A(3)(b) of the Sixth Directive – or is to be treated (wholly or in part) as attributable to a non-monetary element of the total consideration which has been obtained by the supplier for the supply – and so brought into the computation of the taxable amount under art 11A(1)(a) at an appropriate value.

65. When the question is identified and posed in that form it can be seen that it raises two sub-issues: (i) whether there is any reason to apportion the amount claimed as "commission in goods" so as to distinguish between so much of that amount (equal to 80% of the whole) as would represent commission at the "taken in cash" rate of 10% and the balance of that amount (equal to 20% of the whole) which reflects the additional 2.5% allowed because the commission is "taken in goods"; and (ii) whether there is any reason to distinguish between that part of the amount claimed as "commission in goods" which is attributable (or which can be attributed) to commission earned on payments made by the agent in respect of her own purchases and that part which is attributable (or which can be attributed) to commission earned on payments in respect of third party purchases."

80. It can be seen, therefore, that the Court broadened the issues for determination so as to encompass not only the correct treatment of the whole amount of commission taken in goods (was it a price discount at the time of supply, or was it to any extent attributable to a non-monetary element of the consideration for the supply?), but also the question whether there was any reason to apportion such commission between the 10% and 2.5% elements, and the further question whether there was any reason to distinguish between AOP commission and TPP commission. In other words, the Court considered it necessary to extend the scope of the enquiry beyond the 2.5% element which was actually in dispute, and to examine the rationale for splitting commission taken in goods, whether by reference to the 10% and 2.5% elements, or by reference to the source (AOPs or TPPs) from which it was derived.
81. Having reformulated the issues, the Court then made some significant observations about the factual context in which they arose:

"66. It is, we think, important to keep in mind the factual context in which those issues arise. That factual context includes the following elements: (i) there is no direct link between the placing of an order and the credit of commission to the agent's commission account – commission is credited when payments are received, not when orders are placed; (ii) an agent earns commission on all payments made by her to Littlewoods – whether those payments are made by the agent out of her own monies in respect of purchases for her own use or out of monies which she has collected in respect of sales to third parties; (iii) as between Littlewoods and the agent there is no distinction between payments made in respect of purchases for her own use and purchases by third parties – the right to commission, the amount of the commission and the option to take the commission "in goods" or "in cash" is the same whether the payment is made in respect of her own purchase or in respect of a purchase by a third party; and (iv) unless and until the agent chooses to take the commission – whether "in goods" or "in cash" – the treatment for the purposes of VAT is the same whether the commission is attributable to a payment made in respect of her own purchase or in respect of a purchase by a third party; in either case the taxable amount on the goods supplied is the full catalogue price and no reduction is allowed at the time when the commission is credited to her account."

Elements (i) to (iii) of the factual context as described by the Court were firmly grounded in the undisputed underlying facts, and they underpin much of the Court's reasoning. Element (iv) is perhaps more part of the legal than the factual context. It reflects the timing point which the ECJ had recently determined in Freemans.

82. The core of the Court's reasoning is contained at [67] to [75]. The Court began by picking up the timing point:

"67. It is only when the agent chooses to take the commission that a difference in treatment can arise. If she takes the commission "in cash" then – at the least in so far as the commission is attributable to payments made in respect of her own purchases – a reduction is allowed under art 11C(1) of the Sixth Directive. That was decided by the Court of Justice in the Freemans case. The decision in the Freemans case is silent as to the position where the commission taken "in cash" is attributable to payments made in respect of purchases by third parties. But (without deciding the point) we find it difficult to see why art 11C(1) should not be engaged in such a case also. The principle of neutrality, upon which the court relied in reaching its decision in the Elida Gibbs case, points to that conclusion."

This paragraph is relied upon by Littlewoods for the Court's firmly expressed, but admittedly obiter, view that where TPP commission is taken in cash, it should be treated as a discount within article 11C(1), in the same way as AOP commission which is taken in cash. The latter point, as I have explained, has always been common ground; and the decision in Freemans was concerned only with timing. It is therefore not immediately obvious, at least to me, why the Court seemed to think it followed that TPP commission taken in cash should be treated in the same way. In particular, as Mr Swift observed in oral argument, this conclusion appears to pre-empt any consideration of the question whether TPP commission should be regarded as consideration for the agent's services.

83. The Court then turned to the question of apportionment of commission taken in goods, and began its analysis as follows:

"68. The basis of assessment for which the Commissioners have contended – and which was upheld by the judge – does involve an apportionment of the amount claimed as "commission in goods" so as to distinguish between the part (80%) equivalent to commission at the rate of 10% and the part (20%) equivalent to the additional 2.5% allowed because the commission is taken "in goods". But it can now be seen, in the light of the decision of the Court of Justice in the Freemans case, that such a distinction cannot be supported. The true analysis of the position is this: (i) the agent's right to commission (whether "taken in goods" or "taken in cash") cannot be taken into account under art 11A(3)(b) of the Sixth Directive in order to reduce the taxable amount of goods ("primary goods") the sale of which (whether for the agent's own use or to a third party) has, upon payment, generated a credit to the agent's commission account; (ii) when the agent withdraws commission from her commission account "in cash", the withdrawal is treated then – at the least in so far as the commission is attributable to payments made in respect of her own purchases – as a price reduction in respect of the primary goods made "after the supply takes place" and a corresponding allowance is made, at that stage, under art 11C(1); (iii) where the agent takes her commission "in goods" there is no basis for an allowance against the taxable amount of the primary goods under art 11C(1) – unless commission taken in the form of goods ("secondary goods") can be regarded as a post-supply reduction in the price of the primary goods, which is a question which we do not need to address in this context; (iv) if commission taken "in goods" has no effect on the taxable amount of the primary goods, the relevant enquiry – and the only relevant enquiry – in such a case is as to the taxable amount to be attributed to the supply of the secondary goods; (v) in that context, the question is whether the difference between the catalogue price of the secondary goods and the monetary amount (if any) paid by, or debited to, the agent is to be treated as a price discount allowed and accounted for at the time of the supply within art 11(A)(3)(b) of the Sixth Directive or is properly to be treated as attributable to a non-monetary element of the total consideration; and (vi) in addressing that question there is no basis for a distinction between so much of the commission "taken in goods" as is equivalent to commission at the rate of 10% and so much of that commission as is equivalent to the additional 2.5%.

69. Not only does the basis of assessment which the Commissioners seek to adopt involve an apportionment of the amount claimed as "commission in goods" into two parts – the one (80%) equivalent to commission at the rate of 10% and the other (20%) equivalent to the additional 2.5% - it also involves a further apportionment so as to distinguish between so much of the amount claimed as is attributable to commission earned on payments made in respect of the agent's own purchases and so much as is attributable to commission earned on payments in respect of third party purchases. If the basis of assessment which the commissioners seek to adopt is to be supported, that is a necessary distinction. It is necessary because there is no basis that we can identify – and none has been suggested – upon which so much of the amount claimed as "commission in goods" as is attributable to commission earned on payments made by the agent in respect of her own purchases could be treated as a non-monetary element of the consideration obtained by the supplier for the supply of goods supplied in response to a "Commission in Goods 12½%" form. The agent is entitled to "commission" on payments made in respect of her own purchases – and is entitled to take that commission as "commission in goods" – whether or not she has done anything in relation to third party purchases. It seems to us beyond argument that in so far as the difference between the catalogue price for secondary goods ordered and supplied is attributable to commission on payments made by the agent in respect of her own purchases the amount of that difference must be regarded as a price discount or rebate allowed to the customer and accounted for at the time of supply within art 11A(3)(b) of the Sixth Directive. The position is indistinguishable from that in the Boots case.

70. It follows that, unless there is some reason to distinguish between that part of the amount claimed as "commission in goods" which is attributable to commission earned on payments made by the agent in respect of her own purchases and that part which is attributable to commission earned on payments in respect of third party purchases, the whole of the amount claimed as "commission in goods" ought to be treated as a price discount or rebate within art 11A(3)(b)."

84. I would make the following comments on the lengthy passage which I have just quoted. First, the references at the beginning of [68] and [69] to the basis of assessment contended for by the Revenue are puzzling, and must I think have been influenced by the erroneous formulation of the dispute between the parties in [58]. The Revenue's contention, which Lightman J had upheld, was that the whole of the 12.5% commission in goods should be treated as consideration for services, where the commission was attributable to third party purchases. It was the previously agreed treatment of such commission which involved an apportionment, but the Revenue's case was now that this treatment was mistaken. Furthermore, it had never been contended on either side that commission taken in goods should be apportioned where it was attributable to agents' own purchases. It has always been common ground that the whole of such commission should be treated as a discount: see paragraph 32 above. Secondly, I am equally puzzled by the Court's reliance on Freemans in the second sentence of [68]. The timing issue in that case appears to me to have nothing to do with apportionment. That said, however, and this is my third point, the rather densely worded analysis of the "true position" which follows seems to me either to repeat well-established principles which were not in dispute (e.g. as to the treatment of commission generated by AOPs) or to reject the case advanced by Littlewoods (not the Revenue) that a distinction should be drawn between the 10% and 2.5% elements of commission derived from TPPs.
85. The central point which emerges from the discussion, in my view, is that there is no legal or logical reason for such an apportionment, and the whole (not part) of the TPP commission must be treated as a discount at the point of sale of the secondary goods to the agent unless it is right to regard such commission as the value attributed by the parties to a non-monetary element of the total consideration for the sale, i.e. as a reward for the agent's services.
86. That is the question which the Court then proceeded to address:

"71. The reason advanced in support of a distinction between that part of the amount claimed as "commission in goods" which is attributable to the agent's own purchases and that part which is attributable to third party purchases is that, in the latter case (as the judge found), there is the necessary direct causative and contractual link between the agent's services in relation to third party sales and the supply to her of secondary goods in return for a monetary payment which is less than the full catalogue price. It is said that, in such a case, the difference between the full catalogue price for the secondary goods and the monetary payment is attributable (in part at least) to a non-monetary element of consideration, provided by the agent in the form of services. As we have already pointed out, if that is the correct analysis, it applies as much to the whole of the amount claimed as "commission in goods" – so far as attributable to third party purchases – as it does to the "enhanced rate" element of that commission. There is no distinction to be drawn, in this context, between commission at the "in cash" rate of 10% and commission at the "in goods" rate of 12.5%. The only relevant commission is commission at the "in goods" rate.

72. In our view the judge was wrong to find that the necessary direct link between the service provided by the agent in respect of third party sales and the consideration received by Littlewoods in respect of the supply to her of secondary goods had been established. It may be true, as the judge observed, in the passage to which we have already referred (see [2002] STC 588 at 591, para 9), that:

"The reason and only reason why the agent could and did enter into the sale contract at the price reflecting the special allowance [of an additional 2.5%] was the existence and exercise by the agent of the third option contained in the agency contract".

But it is important not to overlook (i) that entry into the agency agreement imposed no obligations upon the agent, (ii) that commission was not earned on sales, but was generated by payments, (iii) that, as between Littlewoods and the agent, it was immaterial whether those payments were made in respect of the agent's own purchases or in respect of sales to third parties, and immaterial whether those payments were made out of the agent's own monies or out of monies collected from third party purchasers, (iv) that the agent could enter into "the sale contract" – that is to say, the contract for the supply of secondary goods – "at the price reflecting the special allowances" whether or not she had provided any service in relation to sales to third parties.

73. Upon a true analysis the agent's right to take commission "in goods" at the rate of 12.5% arose from a combination of two factors: (a) her appointment as an agent and (b) payments made by her in respect of the supply of primary goods. There was no direct link between her right to take commission "in goods" at the rate of 12.5% and any service which she had, or had not, provided in relation to a third party sale. The position is not the same as that in either the Naturally Yours case or the Rosgill case. The decisions in those cases do not provide the answer in the present case.

74. The point may be illustrated by an example. Suppose a case in which the agent makes purchases for her own use to a catalogue value of £200, and makes purchases for third parties to a catalogue value of £100. The amount debited to her account will be £300. She makes payments to the credit of her account which total £150. She is credited with commission which is expressed as "£18.75 if taken in goods, £15 if taken in cash". She chooses to take her commission partly in goods and partly in cash. The catalogue price of the secondary goods which she orders on a "Commission in Goods 12½%" form is £6.25. On a "Commission in Cash 10%" form she requests that £10 be applied towards the discharge of the balance (then £150) on her account. Those dispositions exhaust her commission. In the circumstances that she has made no monetary payment for the secondary goods supplied to her the difference between the catalogue price of those goods and the monetary payment is equal to the amount of the commission "taken in goods" – that is to say, £6.25. But there is no basis upon which it can be said that there is any direct link between that amount and the sales to third parties. There are at least two points at which the link fails: (i) it is not established that any part (or, if so, what part) of the £150 which she has paid to the credit of her account is to be treated as payment for the third party purchases; and (ii), even if it were established that that sum was to be applied, first, in paying the amounts owing in respect of third party purchases (£100), it is not established that any part of the commission "taken in goods" (£6.25) is attributable to commission earned on payments (£100) made in respect of third party purchases.

75. In the absence of any direct link between (a) the difference between the full catalogue price of the secondary goods supplied to the agent and the monetary amount (if any) which she pays for those goods and (b) services provided by her as agent in connection with the sale of primary goods to third parties, there is no basis for treating the provision of services as a non-monetary element in the consideration for the supply of the secondary goods; and no basis for taking a non-monetary element into account (at an appropriate value) under art 11A(1)(a) of the Sixth Directive when ascertaining the taxable amount to be attributed to that supply. In those circumstances the commission "in goods" must be treated as a price discount or rebate allowed to the agent at the time of the supply of the secondary goods; and cannot be included in the taxable amount (see art 11A(3)(b)).

76. For those reasons we allow Littlewoods' appeal."

87. It can be seen from this passage that the appeal ultimately turned, in the view of the Court, on the impossibility of establishing the necessary direct link between the services provided by the agent in respect of third party sales and the consideration received by Littlewoods in respect of the supply to her of secondary goods: see the first sentence of [72]. The Court reached this conclusion, it is important to note, even though the commission in question was, ex hypothesi, commission derived from third party sales. Despite that fact, the Court clearly considered that the "direct link" test was not satisfied. The reason why it was not satisfied was the impossibility of demonstrating, in relation to any particular purchase of secondary goods, that the commission so applied did in fact derive from third party purchases. This was then illustrated by the example in [74].
88. The Court clearly considered that the normal impossibility of establishing a direct link with individual purchases of secondary goods was fatal, even though:

(a) one can readily imagine particular cases where the link would have been indisputable (e.g. where the agent had never made any own purchases, or had not done so since her commission account last had a nil balance);

(b) the records which she kept at the time, and submitted to the company on a monthly basis, would have distinguished clearly between purchases which she made on her own account and those made for third parties; and

(c) at a global level, the company could, and did, distinguish between commission generated by TPPs and that generated by AOPs, not least because it was necessary to do so in view of the different VAT treatment accorded to each category.

89. The result, surprising though it may seem at first sight, was that none of the globally ascertained commission attributable to third party purchases could be treated as consideration for the services of agents in relation to those purchases, because of the practical impossibility (in most, but not all, cases) of demonstrating the necessary link between (a) the commission actually applied by an agent in making a purchase of secondary goods, and (b) third party purchases which had generated that commission. Furthermore, it is implicit in the Court's reasoning, in my judgment, that had the necessary direct link been established, Littlewoods' appeal would have been dismissed. There would then have been no answer to the simple argument which had all along underpinned the agreed treatment of the 10% element of commission, whether taken in cash or goods, namely that it represented consideration for the agent's services in relation to third party purchases.
90. With the greatest respect to the full and careful reasoning of the Court of Appeal, I cannot say that I find its conclusion an altogether satisfactory one. It appears to turn on the impossibility of establishing, at the micro level, a test which Littlewoods and the Revenue had no difficulty in applying at the macro level. On any view, the agents did provide services of real economic value to Littlewoods in procuring, administering and making payments in respect of third party purchases. Nor can it be disputed that, when such payments were made to Littlewoods, they generated a right to commission, which the agent could take either in cash or in goods. Why, then, when the commission was taken in goods, should the direct link to third party purchases be treated as broken? It may not have been demonstrable on a case by case basis, but on an aggregated basis the link was surely just as real and direct as it would have been in a simple case where the agent had never made any purchases on her own account. The parties were content to use an aggregated basis to divide AOPs from TPPs. It seems odd, therefore, to deny the possibility of applying such an approach to purchases of secondary goods viewed collectively, merely because the necessary link could not be demonstrated in respect of each individual purchase. Furthermore, if such an objection were truly fatal, would it not also invalidate the agreed treatment of commission generated by AOPs? It is common ground that such commission is correctly treated as a discount, whether taken in cash or in goods. But if one were to examine the circumstances of each individual purchase of secondary goods, it would usually be just as impossible to demonstrate that the commission in question derived exclusively from AOPs as it would be to demonstrate that it derived exclusively from TPPs.
91. There is one further point which I would make about paragraph [72] of the judgment. The first of the four matters which the Court says it is important not to overlook is "that entry into the agency agreement imposed no obligations upon the agent". As an observation about the arrangements entered into between the company and the agent, that is undoubtedly true. The agent was under no contractual obligation to seek out third party purchasers. But the observation is irrelevant to the question whether the agent performed services of real value to Littlewoods if and when she did introduce third party purchasers. I mention this point, because it was picked up by Advocate General Kokott in Grattan (ECJ), to which I will now turn.

(2) Grattan (ECJ)

92. The background to the reference to the ECJ in the Grattan case was described as follows by the FTT (Judge Berner and Mr Stafford) in Grattan Plc v Revenue and Customs Commissioners [2011] UKFTT 31 (TC), [2011] SFTD 297 ("Grattan (FTT)"):

"3. Grattan is the representative member of a VAT group whose members carry on the business of retailing goods by mail order. In the course of that business companies that are now within the Grattan group used the services of persons described as "agents". Those agents earned commissions (or credits) in relation both to their own purchases of goods from the mail order catalogue (agents' own purchases or "AOP") and purchases by third parties, (third-party purchases or "3PP"). That commission, which was credited to the agents' accounts with the relevant mail order company, was calculated as a percentage of the payments received by the relevant mail order company in respect of the relevant purchases, and could be taken in goods or in cash.

4. According to evidence of Mrs Patricia Vann, a consultant to Grattan who was Grattan's indirect tax manager from April 2001 to July 2008, which was not contested, in a traditional agency situation the agent would hold one account and have a limited number of third-party "sub-customers" to whom she would pass the catalogues. The agent would place the sub-customers' orders by telephoning the call centre or by sending off an order form. Unless an alternative delivery address was specified, the goods ordered would be delivered to the agent, for onward distribution to sub-customers. Payment for goods, usually in instalments, would be collected by the agent from the sub-customers and periodically remitted to the mail order company.

5. We heard that the Grattan and Freemans brands (both within the Grattan group) operated different customer systems. In a traditional agency situation, the Grattan brands would not know who the sub-customers were. However, the more complex customer system operated by the Freemans brands identified the sub-customers and held details of their orders and payments. When Freemans joined the Grattan group (on 27 February 2000), both systems continued to operate in tandem.

6. The agents earned commission in relation to the goods they bought for themselves and the goods purchased by their sub-customers. Goods were generally paid for in instalments and agents were provided with statements every 28 days. For the Grattan brand, agents would be issued with collection cards enabling them to keep their own records of sub-customers' orders and payments which could then be tallied up and compared with the statements received from Grattan.

7. The agents earned commission of 10% based on the amounts paid by themselves and their sub-customers. The commission would be credited to an account and the agent could then claim it as a cheque payment, as a credit against their account balance or as a full or part payment against the purchase of further goods ("secondary goods"). One of the brands within Grattan, "Look Again", provided a further 2.5% commission if the agent used the commission to purchase further goods."

93. As can be seen from the above summary, the underlying facts in Grattan were very similar to those of the present case. The only significant difference appears to have been that, where an agent took commission in goods, there was normally no enhancement of the rate from 10% to 12.5%, although such an enhancement was provided for the "Look Again" brand of goods.
94. It is important to stress at the outset that the dispute in Grattan related only to 10% commission in cash paid or credited to agents, to the extent that such commission was attributable to third party purchases. The question was whether, when TPP commission was taken in cash, it should be treated as reducing the taxable consideration for previous own purchases made by the agent. The case was not concerned in any way with the VAT treatment of the TPPs from which the relevant commission derived. Nor was it concerned with commission taken in cash which was itself attributable to earlier AOPs, or with commission (whatever its source) which was taken in goods. The focus was therefore solely on the issue whether the taxable consideration for AOPs was retrospectively reduced by TPP commission which the agent took in cash.
95. As in the present case, such commission was originally agreed to be consideration for a supply of services by the agent, and therefore not a discount which reduced the consideration for, or taxable amount of, the relevant AOPs: see Grattan (FTT) at [8]. This treatment was later challenged by the companies, and claims were made to recover the VAT allegedly overpaid. Following the decision of the ECJ in Marks and Spencer, these claims were extended back to 1973. In light of the decisions in Freemans and Littlewoods (CA), HMRC accepted that TPP commission taken in cash should be treated as a discount within article 11C(1) of the Sixth Directive, and the relevant tax was repaid together with simple interest under section 78 of VATA 1994. Two matters, however, remained in issue: first, the position under the Second Directive in respect of supplies made before 1 January 1978; and secondly, whether the overpaid tax should be repaid with compound interest.
96. It is convenient at this point to set out the relevant provisions of the Second Directive. Article 2 provided that:

"The following shall be subject to the Value Added Tax:

(a) The supply of goods and the provision of services within the territory of the country by a taxable person against payment;

(b) …"

Article 8 then provided that:

"The basis of assessment shall be:

(a) in the case of supply of goods and of the provision of services, everything which makes up the consideration for the supply of the goods or the provision of services, including all expenses and taxes except the value added tax itself;

…"

97. Article 8(a) was supplemented by paragraph 13 of Annex A to the Second Directive, which, so far as material, provided as follows:

"Regarding Article 8(a)

The expression "consideration" means everything received in return for the supply of goods or the provision of services, including incidental expenses (packing, transport, insurance, etc) that is to say not only the cash amounts charged, but also, for example, the value of the goods received in exchange or, in the case of goods or services supplied by order of a public authority, the amount of compensation received.

…"

98. After a five day hearing in December 2010, the FTT released its decision on 12 January 2011. In short, it decided on the first issue that the position was unclear and the question should therefore be referred to the ECJ for a preliminary ruling, but that determination of the second issue relating to compound interest should be stayed pending the decision of the ECJ in the present case: see Grattan (FTT) at [37] and [70]. On the first issue, the FTT provisionally favoured the arguments advanced by Dr Paul Lasok QC for Grattan to the effect that article 8(a) of the Second Directive should be interpreted in accordance with the principle of fiscal neutrality in such a way as to reduce the taxable amount of the relevant AOP supplies when TPP commission was taken in cash, and that the terms of article 8(a) were unconditional and sufficiently precise to be capable of having direct effect: see [25] to [37].
99. HMRC appealed from this decision to the Upper Tribunal, but on 28 September 2011 their appeal was dismissed by Judge Bishopp and Judge Ghosh QC: see Grattan Plc v Revenue and Customs Commissioners [2011] UKUT 399 (TCC), [2011] STC 2342. The reference to the ECJ on the first issue therefore went ahead. The question referred was framed as follows:

"In relation to the period before 1 January 1978, does a taxable person [i.e. the mail order company] have a directly effective right under art 8(a) of the Second Directive and/or the principles of fiscal neutrality and of equal treatment, to treat the basis of assessment of a supply of goods as retrospectively reduced where, after the time of that supply of goods, the recipient of the supply [i.e. the agent] received a credit from the supplier which the recipient then elected either to take as a payment of money, or as a credit against amounts owed to the supplier in respect of supplies of goods to the recipient that had already taken place?"

100. The case was heard by the Second Chamber of the ECJ, presided over by Judge Rosas. As I have already noted in paragraph 8 above, Advocate General Kokott delivered her opinion on 13 September 2012 and the Court gave its judgment on 19 December 2012.
101. After setting out the relevant provisions of the Second and Sixth Directives, the Court summarised the facts in paragraphs 12 to 15 of the judgment. It noted (paragraph 14) that the dispute before the FTT concerned only situations in which the commission was taken in cash. It also noted (paragraph 15) that HMRC had repaid all of the VAT in issue "with the exception of commission taken in cash in respect of third-party purchase transactions in the period from 1973 up to 1 January 1978". The Court then summarised the parties' contentions before the FTT, and set out the question referred (paragraphs 16 to 18).
102. The Court began its consideration of the question referred by examining article 8(a) of the Second Directive. It concluded that nothing in article 8(a), or anywhere else in the Second Directive, could be interpreted as permitting the retrospective adjustment of the basis of assessment, or of the output tax payable, once delivery of the goods had been effected. The Court's reasoning was as follows:

"19. By its question, the referring tribunal asks, in essence, whether art 8(a) of the Second Directive must be interpreted as conferring upon a taxable person the right to treat the basis of assessment of a supply of goods as retrospectively reduced where, after the time of that supply of goods, an agent received a credit from the supplier which the agent elected to take either as a payment of money or as a credit against amounts owed to the supplier in respect of supplies of goods that had already taken place.

[The Court then referred to the relevant provisions in article 8(a) and paragraph 13 of Annex A of the Second Directive]

22. The expression "consideration" is part of a provision of European Union law which does not refer to the law of the member states for the determining of its meaning and its scope. It follows that the interpretation, in general terms, of the expression may not be left to the discretion of each member state. Such consideration is a subjective value since the basis of assessment for the provision of services is the consideration actually received and not a value assessed according to objective criteria [the Court here referred to the "Dutch Potato" case at paragraphs 9 and 13].

23. In order to determine whether art 8(a) of the Second Directive obliged the member states to permit the consideration to be altered and, therefore, the basis of assessment to be regularised after the chargeable event had taken place, the provisions of that directive concerning the calculation, declaration and payment of VAT should also be examined. Determination of the basis of assessment requires consideration and a chargeable event.

24. It is to be noted in this regard that art 5(5) of the Second Directive provided that "[t]he chargeable event shall occur at the moment when delivery is effected". The term "chargeable event" used in that provision was defined in point 8 of Annex A to the directive as meaning "the event giving rise to the tax".

25. No provision of the Second Directive provided for occurrence of the chargeable event to be set at a subsequent time, or its deferral in another manner. Nor did the directive provide for the alteration of a tax debt that has arisen.

26. Accordingly, as the Advocate General has observed in point 41 of her opinion, pursuant in particular to art 5(5) of the Second Directive a taxable person's tax debt arose in an amount derived from the basis of assessment, a basis which was to be determined at the time of delivery.

27. Therefore, neither art 8(a) nor any other article of the Second Directive could be interpreted as meaning that regularisation of the basis of assessment, or of the output tax, after delivery – which is when the chargeable event took place – had to be permitted."

103. Having reached this conclusion, the Court then considered the principle of fiscal neutrality:

"28. As regards, next, the principle of fiscal neutrality, it is to be noted that this principle, which constitutes a fundamental principle of the common system of VAT, is the reflection in the field of VAT of the principle of equal treatment … One of the consequences of this principle is that taxable persons must not be treated differently in respect of similar supplies which are in competition with each other …

29. The principle of fiscal neutrality is not a rule of primary law which enables on its own the basis of assessment within the meaning of art 8(a) of the Second Directive to be determined (see, to this effect, Finanzamt Steglitz v Zimmermann (Case C-174/11) (15 November 2012, unreported), para 50 and the case law cited). Nor can it make up for the fact that the Second Directive does not include any provision comparable to art 11C(1) of the Sixth Directive.

30. Under the same principle in its other sense, the amount of VAT to be collected by the tax authority must correspond exactly to the amount of VAT declared on the invoice and paid by the final consumer to the taxable person (Fiscale eenheid Koninklijke Ahold NV v Staatssecretaris van Financiën (Case C-484/06), [2009] STC 45, [2008] ECR I-5097, para 36 and the case law cited).

31. It is clear from the documents submitted to the court that, in the main proceedings, sub-customers, as the final consumers of the goods, had to pay the catalogue price for the goods which they purchased and did not receive any commission from the company. The commission was in fact required to be paid back to the agent and not to the sub-customer. In those circumstances, and by virtue of the principles recalled in the preceding paragraph, it must be held that the consideration for the supply corresponded to the full unreduced catalogue price and that the basis of assessment was therefore that price."

104. In this passage, the Court helpfully distinguished between two different senses in which the expression "fiscal neutrality" is used. The first sense (that taxable persons must not be treated differently in respect of similar supplies) was directly relevant to Grattan's arguments, but did not help Grattan because it was "not a rule of primary law" and therefore could not by itself affect the basis of assessment, or make up for the absence in the Second Directive of any provision comparable to article 11C(1) of the Sixth Directive: see paragraph 29. The second sense of fiscal neutrality (that the amount of VAT payable "must correspond exactly to the amount of VAT declared on the invoice and paid by the final consumer to the taxable person") was, however, less obviously relevant, because (as deployed by the Court) it apparently went to a different question, namely whether the subsequent payment of a commission in cash to the agent could in principle reduce the taxable consideration for a previous supply of primary goods to a third party purchaser (the "sub-customer" referred to in paragraph 31). The Court held that no such reduction was possible, because the payment of commission to the agent had no effect on the supply at the full catalogue price to the third party: see paragraph 31.
105. The Court finally considered, and rejected, an argument for Grattan based on the continuity of the VAT system (in paragraphs 32 to 36 of the judgment), before concluding as follows:

"37. Having regard to all the foregoing considerations, the answer to the question referred is that art 8(a) of the Second Directive must be interpreted as not conferring upon a taxable person the right to treat the basis of assessment of a supply of goods as retrospectively reduced where, after the time of that supply of goods, an agent received a credit from the supplier which the agent elected to take either as a payment of money or as a credit against amounts owed to the supplier in respect of supplies of goods that had already taken place."

This conclusion was then reproduced in the formal ruling of the Court.

106. I now turn to the Opinion of Advocate General Kokott. She was clearly concerned that there was a logically prior issue whether the commission in question could in principle reduce the taxable consideration for the relevant AOPs. As she said, in paragraph 26:

"According to the order for reference, the referring tribunal considers that it is only necessary to clarify whether under art 8(a) of the Second Directive a reduction of the basis of assessment is also possible after the time of the supply. However, the first condition for a reduction of the basis of assessment under that provision is that the described payments by the mail order companies are actually repayments of parts of the consideration, which reduce the basis of assessment. This question should therefore be considered first … before I explore whether the Second Directive requires a reduction of the basis of assessment after the supply takes place."

She then said, in paragraph 28, that it was for the FTT to examine whether the payments of commission to the agents constituted repayments of consideration, but in her view the facts set out in the order for reference and the parties' submissions had "raised certain doubts" whether that was the case. In particular, she considered that:

"the legal and business relationships between the mail order companies, the agents and the third-party customers do not appear to be clarified sufficiently and are evidently disputed between the parties to the main proceedings."

107. The Advocate General continued:

"29. Against this background, I consider it necessary to make a few remarks regarding the court's case law to which regard must be had, in order to enable the referring tribunal to give a ruling in the main proceedings which is consistent with EU law on VAT. In so far as the case law cited below does not concern the Second Directive, I consider that it is also applicable to the Second Directive because of the comparability of the provisions of the Sixth Directive interpreted there."

108. The first matter which in her view needed to be clarified was whether the third party sales from which the relevant commission derived were made directly to the third party customers, or whether there was a supply chain with the agents acting as intermediate purchasers. Whatever the position may have been in Grattan, it has never been disputed in the present case that the goods were sold directly by the mail order companies to the customers; and this was indeed the first situation which the Advocate General considered. She did so as follows, in a passage which (because of the significance attached to it by HMRC) I need to quote in full:

"1. Direct sale of goods by the mail order companies to third-party customers

31. If the mail order companies sold directly to third-party customers, as the United Kingdom government has submitted, it would be clear, first of all, that the "commissions" granted to agents for third-party purchases cannot constitute a repayment of the consideration in respect of the sales of goods by the mail order companies to the third-party customers. In that case it would be irrelevant that the mail order companies ultimately received only the catalogue price payable by the third-party customer minus the "commissions" paid to the agents. The court has stated on several occasions that in a situation where a third party is involved in the payment process and, by virtue of this, retains part of the purchase price paid by the purchaser, the full purchase price nevertheless forms the basis of assessment of the supply provided by the seller to the purchaser. The situation can be no different in the present case if the agent, whilst not immediately retaining part of the purchase price paid by the third-party customer, received it subsequently.

32. If the mail order companies sold directly to third-party customers, it is also uncertain whether part of the consideration for another transaction, namely an agent's own purchase, was repaid through the "commissions" granted to the agent for third-party purchases. In principle, this would seem to require a provision to that effect in the legal relationship that defines the consideration for the agent's own purchase. It is true that in Elida Gibbs the court found for the first time that a reduction of the basis of assessment is potentially to be accepted even where the contractually defined consideration is not modified at all. According to that case law, however, at least the payment by the taxable person is conditional on the acquisition of a certain supply by the recipient of the payment. In the main proceedings, it is doubtful whether that condition is satisfied with regard to the basis of assessment of the agent's own purchases, as the credited amounts did not appear to be dependent on an agent's specific own purchase because, in particular, an independent cheque payment was also possible.

33. If a repayment of the consideration for an agent's own purchases were nevertheless to be established, the referring tribunal would then be required to examine whether a reduction of the basis of assessment is ruled out because the "commissions" constitute payment for a supply made by the agents to the mail order companies. If the repayment itself is consideration, there cannot in the final analysis be a reduction of the basis of assessment. For the purposes of determining the basis of assessment, the repaid part of the consideration is replaced by the value of the supply provided. The court has already ruled to this effect in Naturally Yours Cosmetics, in relation to a discount.

34. According to settled case law, a finding that the "commissions" were payment for a supply provided by the agents requires, first, that there was a legal relationship between the agent and the mail order companies pursuant to which there was reciprocal performance. Second, there must be a direct link between the supply and the consideration received. According to the court's case law, that link must be construed as a reciprocal condition of supply and consideration in the context of the legal relationship. In the present case, an agreement should therefore have existed between the mail order companies and the agents, under which payment of "commissions" was made dependent on a supply by the agents. On the other hand, it is not possible to infer further conditions – for example an obligation on the agents to act – from the case law."

109. I would make the following preliminary comments on this passage.
110. First, the reasoning in paragraph 31 appears to be essentially the same as that which the Court adopted in paragraph 31 of its judgment, but (it is important to remember) there was no issue between the parties about the correct VAT treatment of TPPs.
111. Secondly, the Advocate General then moved on to consider two further issues, neither of which was expressly dealt with by the Court in its judgment. The first of those further issues (paragraph 32) was whether commission paid to an agent in respect of third party purchases could ever be treated as a repayment of part of the consideration for her own purchases. This question was of direct relevance to the actual issue in the case, because if TPP commission could not be treated as reducing the consideration for AOPs, that would represent a separate and independent ground for rejecting Grattan's case. It would, moreover, be a ground that on the face of it applied after, as well as before, 1 January 1978. It is notable, however, that the Advocate General's observations on this question in paragraph 32 were couched in tentative and inconclusive language ("it is also uncertain", "this would seem to require", "it is doubtful whether"), quite apart from the fact that they were not expressly endorsed by the Court.
112. The second further issue (paragraph 33), on the assumption that TPP commission taken in cash could in principle reduce the taxable consideration for AOPs, was whether such commission nevertheless constituted payment for a supply of services made by the agent to the company. In that case, as the Advocate General said, "[i]f the repayment itself is consideration, there cannot in the final analysis be a reduction of the basis of assessment". She then went on, in paragraph 34, to set out the requirements for such a finding, including the need for a "direct link between the supply and the consideration received". The authorities cited in support of that requirement were the "Dutch Potato" case at paragraph 12, Naturally Yours at paragraphs 11 and 12, and Lebara Ltd v Revenue and Customs Commissioners (Case C-520/10), [2012] STC 1536, at paragraph 27.
113. In the latter part of paragraph 34, the Advocate General appears to have taken the view that a direct link could be established only if there was a reciprocal agreement in existence between the mail order companies and the agents, under which the payment of commission was made dependent on a supply of services by the agents. On the other hand, it is clear from the final sentence of paragraph 34 that she did not consider it necessary for the agent to be under a contractual obligation to act if the necessary direct link was to be found to exist. In footnote 13, she referred to the judgment of the Court of Appeal in Littlewoods (CA) at [72], where the Court had said it was important not to overlook "that entry into the agency agreement imposed no obligations upon the agent". In her view, therefore, this was not a legally relevant consideration in determining whether a direct link existed.

IV. Are the Woolwich claims and/or the mistake-based claims, as a matter of English law, and without reference to EU law, excluded by sections 78 and 80 of VATA 1994?

114. It is common ground that Vos J has already determined this issue in favour of HMRC. He has decided that all of the claimants' restitutionary claims, whether based on the Woolwich cause of action or the mistake-based cause of action recognised by the House of Lords in DMG, are as a matter of English law excluded by sections 78 and 80 of VATA 1994: see Littlewoods (No. 1) at [45] to [62]. A declaration to that effect is contained in the order for reference to the ECJ made by Vos J on 4 November 2010. By paragraph 3 of that order, Littlewoods' application for permission to appeal from the declaration has been adjourned until after judgment in the two actions. The issue has therefore been finally determined at first instance, and I was not asked to revisit it in any way.
115. It follows, subject only to any future successful appeal by Littlewoods on issue 1, that the present claims must be brought exclusively under EU law. This applies just as much to the underlying issue of liability, which HMRC now wish to re-open, as it does to the questions which were referred by Vos J to the ECJ. Accordingly, issue 2A now asks, in HMRC's formulation:

"As a matter of EU law were the VAT payments in question due in so far as they related to commission (a) taken by the agent by cheque or applied against the balance of her trading account and (b) applied by the agent against the purchase of secondary goods?"

Littlewoods' formulation of the question is identical, save that the order of limbs (a) and (b) is reversed. I cannot see that anything turns on the order in which the two questions are addressed. It may be said that Littlewoods' formulation better reflects the history of the litigation, but as a matter of exposition I find it easier to start with commission derived from TPPs which is taken in cash, so I have adopted HMRC's preferred formulation.

V. The underlying tax issue: as a matter of EU law, were the VAT payments in question due?

116. It is important to begin by defining the scope of the underlying tax issue as closely as possible. As the parties' formulations of the issue make clear, it concerns both commission taken in cash and commission taken in goods. But the issue can be further narrowed down. First, the commission which led to the alleged overpayments of VAT was all TPP commission. There has never been any dispute about the VAT treatment of AOP commission, subject only to the remaining GMAC point in relation to AOP commission taken in cash before 1978. As I have explained that point is not in issue in the present proceedings. Secondly, all of the disputed TPP commission is said by Littlewoods to have been properly allowable as a discount from the price of goods sold to agents, whether as primary goods (i.e. AOPs) or as secondary goods (when the agent exercised the option to take her commission in goods). There has never been any disagreement about the VAT treatment of sales of primary goods to third parties (i.e. TPPs). Thus all of the commission in issue in the present proceedings was TPP commission which allegedly reduced the taxable consideration for sales of primary or secondary goods to agents.
117. It is also worth noting that all the repayments of tax actually made by HMRC have represented the 10% element of the relevant TPP commission, i.e. the whole of the commission if it was taken in cash and the 10% element if it was taken in goods. Tax had of course been paid in respect of the 10% element, because, prior to Littlewoods (CA) and the concession by HMRC of the 10% Commission Appeal, it was thought to represent consideration for the agents' services, and thus could not reduce the taxable consideration for supplies of goods to the agents. On the other hand, the 2.5% element of commission taken in goods had always been thought, before HMRC's change of stance in 1997, to be a discount from the purchase price of the secondary goods, and VAT had accordingly been accounted for on the reduced price. Thus there was no VAT to repay to Littlewoods in respect of the 2.5% element.
118. Turning to the evidence, I believe it now to be common ground (and, if it is not, I find) that, when an agent made a payment which qualified for commission to the relevant mail order company, it was normally impossible for the company to know whether the payment in question derived from TPPs, AOPs or a combination of both. The unchallenged evidence of Mr Mitchell was to this effect, as was the evidence of Mr Maitland which the Court of Appeal had before it in Littlewoods (CA), and which Mr Mitchell has endorsed as accurate: see paragraphs 24 and 30 above. Although Mr Swift appeared at times to wish to cast doubt on this evidence, he accepted at the end of his oral closing submissions that, when an agent took commission from her commission account, there was no distinction between commission which had arisen from AOPs and TPPs, and "there would not have been any particular distinction within the commission account as to the sources of the commission" (transcript, day 12, page 96).
119. The reasons for this were explained by Mr Maitland with great clarity in his witness statement before the Tribunal in 1999:

"25. One issue which is fundamental to our VAT accounting is that it is impossible to reconcile a specific payment to a specific supply. When an agent orders goods the amount she owes is debited to her account. Typically an agent will pay by instalments, and she must make a payment to Littlewoods on a monthly basis to reduce the balance on her account. The amount she is liable to pay in each instalment is calculated by reference to the total amount outstanding. The only way in which we can determine the split between [AOP] and customer purchases is analysing the order forms, and deducting any returned goods to give the net dispatches.

26. Thus each payment could be for goods ordered on the agent's own behalf or for her customers, or a combination of both. The payment, and therefore the commission, may relate to goods dispatched 2 weeks ago or goods dispatched years ago. This is the principle of rolling credit whereby a payment of a proportion of each order is required each month rather than requiring the agent to pay for one item and then another."

120. Despite the impossibility in most cases of ascertaining the source of commission at the time when it was actually taken by the agent from her commission account, there was never any difficulty in distinguishing, on an aggregated basis, between TPP commission and AOP commission. And, to repeat, it was only the treatment of TPP commission which was in dispute between Littlewoods and the Revenue. Building on this point, and on the observations of Advocate General Kokott in paragraph 32 of her opinion in Grattan (ECJ), the first main argument advanced by Mr Swift on behalf of HMRC is that, as a matter of principle, there is no proper basis for treating TPP commission taken by the agent in cash as a deduction from the taxable consideration for AOPs. The economic reality, he submits, is that the TPP commission taken by the agent in cash was a cost to the company which was referable exclusively to the TPPs which she had procured. It had nothing to do with the entirely separate supply of goods to the agent on her own account. Further or in the alternative, the TPP commission should in any event be regarded as consideration for services supplied by the agent in relation to TPPs. Either way, the result is that it cannot go to reduce the taxable consideration for the agent's own purchases. The TPP commission is referable to a different supply chain from AOPs, and it should in any case be regarded as consideration for services rendered by the agent.
121. The disconnection between TPP commission taken in cash, on the one hand, and AOPs on the other hand, is demonstrated, according to the Revenue, by the fact that an agent's entitlement to TPP commission did not depend in any way on her having made any AOPs in the past, or on her doing so in the future. The commission was generated simply and solely by payments made to the mail order company in respect of TPPs arranged through the agent. By contrast, AOPs, if and when they were made, were the product of an independent decision taken by the agent to purchase goods on her own behalf.
122. In relation to TPP commission taken in goods, HMRC submit that, subject to one significant difference, the position is essentially the same as for commission taken in cash. As with TPP commission taken in cash, the supply of goods to the agent (although here in the form of secondary goods, rather than AOPs) is entirely separate from the supply of goods to third parties which generated the commission. The separate nature of the two types of supply is brought out by the fact that it was always open to the agent to take her commission in cash, instead of in goods. She was never under an obligation to take any part of her commission (whether derived from TPPs or AOPS) in goods, although there was an incentive for her to do so in the form of the extra 2.5% allowance for commission taken in goods. It is this extra 2.5% which, according to HMRC, gives rise to the one material difference from the treatment of commission taken in cash. The realistic way to look at the matter, they argue, is that the extra 2.5% is a cash discount from the catalogue price of the secondary goods, allowed to the agent at the time of supply. Unlike the 10% element, which it is always open to the agent to take in cash, the extra 2.5% is directly linked to the supply of the secondary goods, because it is only given if and when the agent decides to make a purchase of secondary goods. In other words, the position for which HMRC are now arguing in relation to the 2.5% element of commission taken in goods is the same as the previously agreed position from which they resiled in May 1997, and diametrically opposed to the position which they unsuccessfully sought to uphold in Littlewoods (CA), namely that the full 12.5% represented consideration for the agent's services in relation to TPPs.
123. Having given this brief outline of the Revenue's main submissions, I find it helpful to consider next how far, if at all, they are mandated, or at least supported, by Grattan (ECJ). It is important to distinguish here between matters decided by the ECJ itself, which are of course binding on domestic courts at all levels in the United Kingdom and other Member States, and views expressed by the Advocate General in her opinion. Unless endorsed by the Court, such views are not (as I understand it) a source of binding EU law in their own right, although they may carry considerable persuasive weight, and they may also serve to clarify or explain (but not to contradict) the reasoning of the Court.
124. The actual issue in Grattan (ECJ), as I have explained, was a narrow one. The issue which divided the parties was whether, in relation to periods before 1 January 1978, the taxable consideration for AOPs could be treated as retrospectively reduced by TPP commission which the agent took in cash. Resolution of this question principally turned on the relevant provisions of the Second VAT Directive, which did not include any equivalent to article 11C(1) of the Sixth Directive, and on the effect of the principle of fiscal neutrality in either of its two senses. As reformulated by the ECJ in paragraph 19 of its judgment, the issue was treated as one of interpretation of article 8(a) of the Second Directive in circumstances where the basis of assessment of goods supplied by a taxable person (i.e. the mail order company) was retrospectively reduced by commission which the agent elected to take in cash.
125. The first ground on which the Court decided the above issue in HMRC's favour depended on the interpretation which it placed on article 8(a) of the Second Directive. This part of the judgment is irrelevant for present purposes, although it is of course highly material to the remaining GMAC point. The only part of the judgment which is potentially relevant to the underlying tax issue in the present proceedings is the discussion by the Court of fiscal neutrality in paragraphs 28 to 31 of its judgment, which I have already quoted in paragraph 103 above. The Court began by explaining, in paragraphs 28 and 29, that the principle of fiscal neutrality in its first sense could not help Grattan, because it is not a rule of primary EU law and it could not by itself make good the absence in the Second Directive of any provision equivalent to article 11C(1). Here again, the Court's reasoning is of no immediate relevance to the present issue. The next two paragraphs, however, are potentially relevant. In paragraph 30, the Court stated that, under the principle of fiscal neutrality in its second sense, the amount of VAT payable to the tax authority must correspond exactly to the amount of VAT paid by the final consumer to the taxable person. The authority to which the Court referred for this proposition was the case of Ahold at paragraph 36 and the case law there cited, which included Elida Gibbs. So far, I can find no indication that the Court was breaking any new ground, or intending to do anything other than repeat a well-established principle. The Court then went on, in paragraph 31, to apply the principle which it had just enunciated to the supply of goods to third party purchasers, i.e. TPPs, and made the point that the consideration for such supplies at the full catalogue price could not be treated as reduced by the payment of commission to the agent. The third parties who were the final consumers of the relevant supplies remained liable to pay the full invoice price of the goods which they had purchased, and the payment of commission by the vendor company to the agent could make no difference to that simple analysis.
126. At this point, a problem presents itself. The situation considered by the Court in paragraph 31 of its judgment had never been in dispute between the parties, and at first sight it has no obvious bearing on the narrow question which was in dispute, namely whether TPP commission taken in cash by the agent could retrospectively reduce the taxable consideration for earlier AOPs. There is also an important factual distinction between the two situations. In the circumstances posited in paragraph 31, the commission was paid to somebody other than the final consumer of the taxable supply, whereas in the latter type of case the commission was paid to the final consumer of the taxable supply of AOPs, i.e. the agent herself. On the other hand, paragraph 31 did repeat essentially the same point which Advocate General Kokott had made in paragraph 31 of her opinion, at the beginning of her discussion of the wider issues in the context of which the narrow issue in dispute had to be placed.
127. Faced with this conundrum, HMRC submit that, by virtue of its clear (if unstated) endorsement of the views of the Advocate General in paragraph 31 of her opinion, the ECJ must be taken to have also implicitly approved her reasoning in the immediately following paragraph 32, where she doubted whether TPP commission could ever be treated as a repayment of part of the consideration for AOPs. Unlike the situation posited in paragraph 31, this part of the Advocate General's reasoning is of direct relevance to the issue in the present case. Thus, the argument runs, by endorsing her reasoning on the (strictly irrelevant) question discussed in paragraph 31 of her opinion, the Court must have intended at the same time to endorse her reasoning on the linked, and directly relevant, question which she proceeded to discuss in paragraph 32. Only in this way, it is said, can meaningful content be give to what would otherwise be an irrelevant digression by the Court in paragraphs 30 and 31 of its own judgment.
128. The argument is ingenious, but I am unable to accept it. In my judgment it is impossible to infer, from the one point which the Court did clearly endorse, a further implied and unarticulated endorsement by it of a separate point which the Advocate General had discussed in notably tentative and uncertain terms. It is far more natural to infer that the Court was intending only to express its agreement with the clear and undisputed point which the Advocate General had made in her paragraph 31, and to leave open the far more uncertain question which she discussed in paragraph 32. It was necessary for the Court to give some consideration to the principle of fiscal neutrality, if only because it had been raised by the national court in the order for reference; but the Court did not need to express any view on the question discussed by the Advocate General in paragraph 32, despite its potential relevance, because the Court had already given adequate reasons for answering the question referred in HMRC's favour. In my view, all that can safely be deduced from paragraphs 30 and 31 of the Court's judgment is that it was, in effect, saying: "This much, at least, is clear as a consequence of the application of the principle of fiscal neutrality, in its second sense, to the wider factual background of the question referred by the national court."
129. For similar reasons, I find it equally impossible to spell out from the silence of the Court any positive endorsement by it of the views expressed by the Advocate General in paragraphs 33 and 34 of her opinion on the question whether TPP commission taken in cash constituted payment for a supply of services by the agent to the mail order company. Here, however, the Advocate General clearly regarded the relevant law as well settled, and it seems to me that her views are entitled to considerable weight in their own right as a helpful distillation of the relevant principles. In particular, she reiterated the well-established need for a "direct link" between the supply and the consideration received, in the sense of "a reciprocal condition of supply and consideration in the context of the legal relationship". On the other hand, she said it was not a requirement of EU law that any further conditions had to be satisfied, including the existence of an obligation on the part of the agent to act. On this last point, she evidently disagreed with the significance which the Court of Appeal had apparently attached to the absence of any such obligation in Littlewoods (CA) at [72]. (I say "apparently attached", because I think it is probably the case, as Mr Rabinowitz submitted, that the Court of Appeal did not intend this factor to be viewed in isolation from the other three to which it drew attention: it was the overall picture which mattered).
130. I conclude from this analysis that Grattan (ECJ) is very far from being the ground-breaking decision that the Revenue submits it to be. In the context of the issues in the present case, and the question whether the repaid VAT was due in the first place, I consider that the only real support that Grattan (ECJ) provides for the Revenue's arguments lies in:

(a) the tentative views expressed by the Advocate General in paragraph 32 of her opinion; and

(b) her rejection of any perceived need for the agent to be under an obligation to act, before a direct link could be found to exist between the supply of services by an agent and the TPP commission which she received.

Neither of these propositions was in my judgment endorsed by the Court, although I think it is safe to say that the Court would not have disagreed with the second of them, which seems to me obviously correct.

131. Having dealt with Grattan (ECJ), I will now examine in more detail HMRC's first main argument that TPP commission taken in cash cannot reduce the taxable consideration for AOPs. In my view, there are several powerful points to be made in support of this argument:

(a) First, it seems wrong in principle that commission which, by definition, is earned in respect of supplies of primary goods to third parties should be allowable as a deduction from the taxable consideration for entirely separate supplies of goods to agents. The two supply chains are quite separate, as is shown by the fact that the agent's entitlement to TPP commission did not depend in any way on her having made any AOPs.

(b) Secondly, the force of the first point should not be blunted by the fact that it is usually impossible to ascertain the source of commission taken by an agent in cash on any particular occasion. There has never been any difficulty in ascertaining global figures for TPP commission and AOP commission, and it must therefore be right to treat the globally ascertained TPP commission as referable exclusively to TPPs, and as having nothing to do with the supply to agents of AOPs.

(c) Thirdly, it would be wrong to proceed on the tacit assumption that, merely because TPP commission paid to the agent cannot reduce the taxable consideration for the TPPs themselves, it must therefore be allowable as a deduction from the taxable consideration for purchases made by the agent on her own behalf. The right analysis could well be, as Mr Swift submitted, that TPP commission taken in cash did not reduce the taxable consideration for any supplies of goods, and was simply an irrecoverable cost to the mail order companies of their supplies of primary goods to third party purchasers.

(d) Fourthly, although Elida Gibbs shows that VAT cannot be charged on more than the amount paid by the final consumer, where discounts or rebates are granted by the original manufacturer in a chain of supply to the final consumer, that principle should have no application where the original supplier of the goods and the final consumer are not in the same supply chain. To extend the principle of Elida Gibbs beyond the confines of a single supply chain would be unprincipled, because it would allow a deduction for an expense which had nothing to do with the relevant supply of goods.

(e) Finally, the views of Advocate General Kokott in paragraph 32 of her opinion in Grattan (ECJ) are of some persuasive force in favour of this analysis, and it can at least be said that nothing in the judgment of the ECJ evinces any disagreement with them.

132. In answer to these arguments, counsel for Littlewoods rely on the reasoning and conclusions of the Court of Appeal in Littlewoods (CA). They accept, as they must, that the view expressed by the Court in [67], to the effect that TPP commission taken in cash reduced the taxable consideration for AOPs, was avowedly obiter: the Court said, in terms, that it was not deciding the point. They submit, however, that this view fits in seamlessly with the reasoning of the Court, as developed in the previous paragraphs, and they suggest that it represents a small and logical extension of the decision of the ECJ in Freemans, where it was held that AOP commission taken in cash fell to be treated as a retrospective discount under article 11C(1) from the taxable consideration for earlier AOPs. Reliance is also placed on Elida Gibbs and the principle of fiscal neutrality, which the Court of Appeal itself considered to be an indicator pointing to the correct conclusion: see the last sentence of [67].
133. In his oral submissions, Mr Rabinowitz made much of the point that in Elida Gibbs the ECJ had not followed the firmly expressed views of Advocate General Fennelly, who considered that in the context of article 11C(1) there had to be a direct link between the giving of a discount and the relevant taxable transaction, if the discount was to be allowable as a deduction from the taxable amount. This view was clearly rejected by the Court, which in paragraphs 28 to 31 of its judgment affirmed the principle that the taxable amount of the original supply of goods by the manufacturer had to be adjusted whenever the amount received by the manufacturer was subsequently reduced (whether by refunds of the value of money-off coupons to the intermediate retailer, or by the payment of cash-back coupons to the final consumer). This result flows from the principle of fiscal neutrality in its second sense, and does not depend on the existence of any contract between the taxable person (the manufacturer) and either the retailer or the final consumer. What all this shows, submitted Mr Rabinowitz, is that fiscal neutrality is a cardinal feature of VAT as a tax on consumption, and it transcends contractual relations in order to achieve neutrality in the tax position of the supplier.
134. The strength of the principle which the ECJ formulated in Elida Gibbs was reinforced, in Littlewoods' submission, by the later decisions of the Court in two cases: Goldsmiths (Jeweller) Ltd v Customs and Excise Commissioners (Case C-330/95) [1997] ECR I-3801, [1997] STC 1073 ("Goldsmiths"), and Freemans. Mr Rabinowitz drew attention in particular to what the ECJ said in Freemans at paragraphs 31 to 33 of its judgment:

"31. Finally, it must be stated that Article 11C(1) of the Sixth Directive must be interpreted as meaning that, in a sales promotion scheme such as that at issue in the main proceedings, the taxable amount constituted by the full catalogue price must be reduced as soon as the agent withdraws or uses in another way the amount with which her separate account has been credited.

32. Freemans and the Commission submit that Article 11C(1) of the Sixth Directive covers the cases in which the reduction in consideration arises from an amendment to the contract after the time of the supply. That provision is therefore not applicable where, as in the present case, the contractual relations giving rise to the supply provide from the very beginning for the grant of a discount, even though that discount is not actually accounted for until later. In that context, they rely upon paragraph 31 of the judgment in Elida Gibbs, cited above, in which the Court held that that provision refers to the normal case of contractual relations entered into directly between two contracting parties, which are modified subsequently.

33. In that regard, it suffices to state that the wording of Article 11C(1) of the Sixth Directive does not presuppose such a subsequent modification of the contractual relations in order for it to be applicable. In principle, it requires the Member States to reduce the taxable amount whenever, after a transaction has been concluded, part or all of the consideration has not been received by the taxable person (see Case C-330/95 Goldsmiths [1997] ECR I-3801, paragraphs 16, 17 and 18). Moreover, there is no indication that in its judgment in Elida Gibbs, cited above, the Court wished to restrict the scope of application of that provision. On the contrary, it is apparent from the facts of the Elida Gibbs case that there had been no modification of the contractual relations. Nevertheless, the Court held that Article 11C(1) of the Sixth Directive was applicable."

135. It is certainly true that in this passage the Court reaffirmed the principle which it had stated in Elida Gibbs, and made it clear that the principle could apply even where there was no change to the contractual relations which gave rise to the original supply. It needs to be remembered, however, that the commission with which the Court was concerned in Freemans was AOP commission, which was derived from and referable to the agents' own purchases. There was no question of a deduction being sought for commission referable to the supply of different goods to third parties. Moreover, the dispute between the parties in Freemans was purely a timing one. There was no disagreement about the principle that AOP commission should, at some stage, reduce the taxable amount of AOPs. In the light of these factors, I do not consider that the decision of the Court in Freemans significantly advances Littlewoods' case.
136. In summary, Littlewoods' argument is that the principle of fiscal neutrality required the taxable amount of AOPs to be reduced whenever commission was subsequently taken in cash by agents, and it did not matter for this purpose whether the commission so taken was AOP commission (as in Freemans) or TPP commission. The Revenue's riposte to this argument, again shortly stated, is that it overlooks the crucial distinction between AOP commission and TPP commission. The principle formulated by the ECJ in Elida Gibbs relates only to transactions within a single supply chain. The ECJ has never yet had to consider a case where the taxable amount of a supply is sought to be reduced by a subsequent payment to the final consumer which relates to a separate supply of different goods to a different person. To extend the principle in such a way, submits the Revenue, would cut across the conceptual foundations of VAT, and would be both commercially and economically unrealistic.
137. During the oral argument, I was provisionally inclined to think that Littlewoods' arguments on this critical issue should prevail. Quite apart from the strong steer given by the Court of Appeal in Littlewoods (CA) at [67], it seemed to me that the principle of fiscal neutrality as expounded in Elida Gibbs and Freemans was probably potent enough to require the taxable amount of AOPs to be reduced by TPP commission subsequently taken in cash, particularly as the transactions all took place within the context of the contractual arrangements made between the mail order companies and the agents. I was also impressed by the practical impossibility of distinguishing between AOP and TPP commission in individual cases. With the benefit of further time for reflection, however, I have increasingly come to the view that the Revenue's submissions are to be preferred. I can see no satisfactory answer to the central point that TPP commission is referable to a separate supply chain, and as such cannot properly reduce the consideration for AOPs, even though it is paid to the same agents who made the AOPs. For the purposes of the relevant supply of goods, namely the AOPs, it is an extraneous payment. In my judgment the principle of fiscal neutrality as expounded in Elida Gibbs is only capable of transcending contractual relations within the context of a single supply chain, at any rate in the absence of express agreement between the parties to the contrary. It is not suggested, however, that there was any agreement in place between the mail order companies and the agents which expressly provided for the catalogue price of AOPs to be subsequently reduced by TPP commission.
138. I am acutely aware that, in reaching this conclusion on a difficult issue, I am differing from the view expressed by the Court of Appeal in Littlewoods (CA), after a full and careful review of all the authorities. However, the Court deliberately left the point open, and although its view is of course entitled to the greatest respect, it did not constitute a necessary step in the Court's reasoning and is not binding on me. Furthermore, the Court does not appear to me to have focused, at this stage in its judgment, on the critical conceptual distinction between AOP commission and TPP commission. The reason why the Court of Appeal said nothing about this point may have been the practical impossibility, at least in most cases, of distinguishing between the two types of commission at the time when it was taken by the agent. But the fact that the two types of commission were inextricably mingled in the agent's single commission account does not mean, in my view, that they lost their conceptually separate character or that the TPP component of the mixed fund thereby became deductible from the taxable amount of AOPs when it was taken in cash by the agent.
139. The next question is whether I feel sufficiently confident about my conclusion, as a matter of EU law, to hold that it is correct without the need for a further reference to the ECJ. As to that, the position is in my view as follows. There is no existing case in which the ECJ has considered the application of the principle of fiscal neutrality, as expounded by it in Elida Gibbs and subsequent cases, to a discount from, or repayment of part of, the taxable amount of a transaction, where the discount or repayment derives from a supply of different goods to a third party, but is granted to the final consumer of the taxable transaction. I have to bear in mind that the views expressed on this point by a very experienced Advocate General in Grattan (ECJ) were couched in terms of uncertainty, and that the question was left open by the Court.
140. My conclusion also differs from that reached, albeit obiter, by the Court of Appeal in 2001, after full and careful consideration of the decisions of the ECJ in both Elida Gibbs and Freemans. Taking all those considerations into account, I cannot say with a good conscience that I regard the issue as acte clair. Accordingly, subject to the outcome of any appeal from my decision, and on the further assumption that it were necessary for my ultimate decision of the case, I would reluctantly see no alternative to the making of a further reference to the ECJ for a preliminary ruling on the issue.
141. I move on to the Revenue's second main argument in relation to TPP commission taken in cash, namely that it represents consideration for services rendered by the agents. If the question were free from authority binding upon me, I would see much attraction in the simple argument that, as a matter of commercial reality, TPP commission taken in cash should be regarded as consideration for the services of the agent in procuring and administering the sales to third parties from which the commission derived. Indeed, if I am right in my conclusion on the Revenue's first main argument, it seems to follow naturally that TPP commission should be characterised in this way. There can be no doubt about the reality and commercial value to the mail order companies of the services in question, and this was, of course, reflected in the agreed treatment of such commission for VAT purposes until the ruling in Littlewoods (CA).
142. Furthermore, I would be disposed to hold, if I were free to do so, that the EU law requirement of a direct link was satisfied, at what I have termed the global or macro level, and that the impossibility of establishing such a link on a case by case basis, as and when commission was taken by an agent, did not mean that the link was either indirect or non-existent. It was, rather, a direct link which had to be established by looking at the evidence as a whole, and asking whether there was the necessary direct relationship between the globally ascertained figure for TPP commission on the one hand, and the services provided by the agents in relation to third party sales on the other hand.
143. Whatever may be the theoretical merits of such an argument, however, I am satisfied that for a judge sitting at first instance this part of the case must be governed by the decision of the Court of Appeal in Littlewoods (CA). It will be remembered that, in [70] to [75] of its judgment, the Court considered, and rejected, the contention that either the whole or some part of TPP commission taken in goods should be treated as consideration for the agent's services. Reversing Lightman J, the Court held that the necessary direct link had not been established. Having regard to the four factors identified in [72], the Court said that "upon a true analysis" the agent's right to take commission in goods at the rate of 12.5% arose from a combination of only two factors, namely her appointment as an agent and payments made by her in respect of the supply of primary goods. This was then illustrated by the detailed example given in [74]. Speaking for myself, I do not find the reasoning of the Court on this issue particularly persuasive, but that is not the point. It clearly forms part of the ratio of the decision, so far as concerns TPP commission taken in goods, and to that extent is formally binding upon me. The relevant facts are, in the case of the Littlewoods claimants, identical, and in the case of the GUS claimants, materially indistinguishable. Further, the reasoning of the Court is clearly just as applicable to TPP commission taken in cash as it is to TPP commission taken in goods. There is therefore no principled basis upon which I could loyally follow the Court of Appeal's decision in relation to TPP commission taken in goods, but feel free to depart from it in relation to such commission taken in cash.
144. The position would be different if it were the case that a subsequent decision of the ECJ had shown the Court of Appeal's reasoning on this issue to be erroneous. I invited counsel to address me on what approach I should adopt in that hypothetical state of affairs, and they were able to satisfy me, by reference to such authority as there is on the subject, that it would probably be my duty in those circumstances to follow the later decision of the ECJ. That is, indeed, what I would have expected the position to be. But if I am right in my analysis of Grattan (ECJ), the question simply does not arise. The ECJ said nothing on the consideration issue, and the well-established principles to which the Advocate General referred in her opinion were in substance no different from those which the Court of Appeal considered and applied in Littlewoods (CA).
145. In these circumstances, the only point upon which HMRC can fasten is the disapproval by the Advocate General in Grattan (ECJ) of the first of the four factors identified by the Court of Appeal in paragraph [72] of its judgment. In my view, however, there is nothing in this point. Quite apart from the fact that the Advocate General's observations were not expressly endorsed by the Court, the most that could be said is that, taken in isolation, one of the four factors identified by the Court of Appeal was strictly speaking irrelevant to establishment of the necessary direct link. As I have already said, I think it is probably wrong to look at the four factors disjunctively, and it was the picture as a whole which mattered. As an observation on the facts, the first factor was clearly correct. Furthermore, I am not satisfied that it was irrelevant. Had it been the case that agents were under a contractual obligation to procure third party sales, that would no doubt have been a feature which told strongly in favour of the necessary direct link. It was therefore material to note that there was no such obligation, without in any way implying that its absence was in itself incompatible with the existence of a direct link.
146. Finally, I come to the treatment of TPP commission taken in goods. In the light of what I have said about Grattan (ECJ), I can be brief. I am bound by Littlewoods (CA) to hold that the whole 12.5% of TPP commission taken in goods:

(a) represents a discount allowed to the agent at the time of supply from the catalogue price of the secondary goods, within article 11A(3)(b) of the Sixth Directive; and

(b) in the absence of the necessary direct link, none of it is to be treated as consideration for the agent's services in relation to TPPs.

147. If, however, the question were free from domestic authority which is binding on me – which it is not – I would hold, for the reasons which I have already given in relation to TPP commission taken in cash, that:

(a) the question whether the 10% element of the commission could be allowed as a discount from the price of secondary goods supplied to the agent should be referred to the ECJ, my own view being that it could not (because it is exclusively referable to TPPs, just like the 10% TPP commission taken in cash);

(b) the 2.5% element of the commission, by contrast, is rightly to be regarded as a discount allowed to the agent at the time of supply of the secondary goods (in accordance with the settled practice before the Revenue's change of mind in May 1997); and

(c) the 10% element is in any event to be treated as consideration for services rendered by the agent in relation to TPPs, in the same way as TPP commission taken in cash.

148. I would add that, even if the 2.5% were not deductible as a discount, it would in practice make no difference to the outcome of the present case, because (first) all of the tax originally paid to the Revenue, and subsequently repaid by it, relates to the 10% element, and (secondly) the Revenue's new arguments on liability are put forward only as a defence to the present claims.
149. I conclude my judgment on this part of the case with the following summary.

(1) Grattan (ECJ) does not in my judgment show that any of the VAT in issue in the present case was in fact lawfully due when it was originally paid or accounted for to the Revenue.

(2) Sitting at first instance, I am therefore bound by Littlewoods (CA) to hold that the whole of the relevant TPP commission taken in goods represented a discount from the price of the secondary goods, and that none of it represented consideration for the agent's services. Accordingly, all of the VAT paid or accounted for on the footing that the 10% element of TPP commission taken in goods was not deductible from the taxable amount of the secondary goods was not due and was overpaid.

(3) In relation to TPP commission taken in cash, the question whether such commission represented a retrospective discount from the taxable amount of earlier AOPs is uncertain as a matter of EU law, although my own view is that it did not. The question would therefore need to be referred to the ECJ, if its resolution were necessary for the proper disposal of the present case. If the ECJ were to agree with my provisional view, the result would be that all of the VAT originally paid or accounted for on the relevant AOPs was lawfully due, and it remained so when the commission was taken, there being no scope for article 11C(1) of the Sixth Directive to operate.

(4) I am bound by Littlewoods (CA) to hold that TPP commission taken in cash did not represent consideration for the agent's services in relation to TPPs. This conclusion does not, however, impact on paragraph (3) above, because it is enough for the Revenue to succeed on either of their two main arguments in order to show that the VAT in question was originally due.

(5) If I were not bound by Littlewoods (CA), my inclination would be to accept the Revenue's submission that the whole of TPP commission taken in cash, and the 10% element of TPP commission taken in goods, represented consideration for the agent's services, with the consequence that all of the VAT in dispute in the present claims was lawfully due when it was originally paid or accounted for.

VI. Estoppel and abuse of process

(1) Introduction

150. The parties are agreed on the formulation of issue 2B, which reads as follows:

"2B. If the VAT payments in question, or some of them, were due, are [HMRC] estopped from defending the claims, or some of them, on this basis? Is it otherwise an abuse of process for [HMRC] to do so?"

In very general terms, therefore, the questions which I have to consider under this heading ask whether (and, if so, to what extent) it is open to HMRC to contend that the relevant payments of VAT were due in the first place.

151. Since I have already examined the underlying issue of liability, and stated my conclusions on it, in the preceding section of this judgment, it may at first sight seem paradoxical that I should only now consider the logically anterior question whether HMRC should be permitted to argue the underlying issue at all. Both sides were in agreement, however, that the issues should be argued and dealt with in this order, given the unusual procedural history of the case which has led to both issues coming before the court for determination at the same time. In the usual way, the questions of estoppel and abuse of process would have been resolved on HMRC's application to re-amend their defences to the claims; and that was, of course, the original intention at the abortive hearing before Vos J in April 2013 (see paragraphs 8 to 10 above). Once the court had decided to postpone the resolution of those questions until trial, however, there was no longer anything to be gained by treating them as preliminary issues. In the interests of clarity, it is far easier to deal with those questions after, rather than before, discussing the underlying liability issue. A further reason for doing so is that, in relation to some of the arguments on issue estoppel and abuse of process, it is material for the court to consider the strength of HMRC's case on the underlying issue. It therefore makes obvious sense for the court to determine the underlying issue before forming a view on the strength of HMRC's case.

(2) Issue estoppel

(a) General requirements

152. Issue estoppel is a well-established part of the law of res judicata. It is common ground that, in order for an issue estoppel to arise, three conditions need to be satisfied:

(i) the same question must previously have been decided;

(ii) the judicial decision which is said to create the estoppel must have been a final decision of a court of competent jurisdiction; and

(iii) the parties to the prior judicial decision (or their privies) must have been the same persons as the parties to the subsequent proceedings in which the estoppel is raised (or their privies).

See, for example, Carl Zeiss Stiftung v Rayner & Keeler Ltd (No. 2) [1967] 1 AC 853 (HL) at 935A-C per Lord Guest (on which the above formulation is largely based) and the well-known exposition of the subject by Diplock LJ in Thoday v Thoday [1964] P 181 at 197-198.

153. In Arnold v NatWest Bank Plc [1991] 2 AC 93 ("Arnold"), Lord Keith of Kinkel distinguished between cause of action estoppel (at 104D-E) and issue estoppel (at 105D-E), as follows:

"Cause of action estoppel arises where the cause of action in the later proceedings is identical to that in the earlier proceedings, the latter having been between the same parties or their privies and having involved the same subject matter. In such a case the bar is absolute in relation to all points decided unless fraud or collusion is alleged, such as to justify setting aside the earlier judgment. The discovery of new factual matter which could not have been found out by reasonable diligence for use in the earlier proceedings does not, according to the law of England, permit the latter to be re-opened.

…

Issue estoppel may arise where a particular issue forming a necessary ingredient in a cause of action has been litigated and decided and in subsequent proceedings between the same parties involving a different cause of action to which the same issue is relevant one of the parties seeks to re-open that issue."

154. In Virgin Atlantic Airways Ltd v Zodiac Seats UK Ltd [2013] UKSC 46, [2014] AC 160 ("Virgin Atlantic"), Lord Sumption JSC has recently summarised the law on res judicata in terms with which the other four members of the Supreme Court all agreed: see [17] to [26], and [42]. At [20] Lord Sumption endorsed the restatement by Lord Keith in Arnold, quoted above, of "the classic distinction" between cause of action estoppel and issue estoppel. In [17], Lord Sumption himself encapsulated the distinction in similar terms, describing issue estoppel as:

"… the principle that even where the cause of action is not the same in the later action as it was in the earlier one, some issue which is necessarily common to both was decided on the earlier occasion and is binding on the parties: Duchess of Kingston's Case (1776) 20 State Tr 355."

155. In the present case, there are technical issues between the parties about the extent to which the first and third of the general requirements set out in paragraph 152 above are satisfied, if one reaches the stage that the doctrine of issue estoppel is capable of applying at all to the circumstances of the case. There is no dispute, however, about the second requirement. In other words, it is agreed that the earlier decisions upon which Littlewoods rely (i.e. the actual decision of the Court of Appeal in Littlewoods (CA), and the determinations by the Tribunal and the FTT respectively which are deemed to exist by virtue of the 2004 and 2008 section 85 Agreements) have the necessary qualities of finality and jurisdictional competence.
156. Before coming on to those technical issues, I will first consider two overarching contentions which Mr Swift placed at the forefront of his oral submissions for HMRC on this part of the case. The first contention was that HMRC should not be prevented from relying on Grattan (ECJ), because such reliance would accord with the public policy rationale of issue estoppel, and would fall within the exception to issue estoppel recognised by the House of Lords in Arnold and affirmed by the Supreme Court in Virgin Atlantic. The second contention was that issue estoppel in any event has only very limited scope for operation in tax cases, both as a general proposition (largely based on what is sometimes called "the tax exception", as recognised by the Privy Council in Caffoor v Income Tax Commissioner [1961] AC 584 ("Caffoor") and subsequent case law), but also with particular reference to the facts that the present claims (a) relate exclusively to interest, and (b) concern VAT.

(b) Does the present case fall within the Arnold exception to issue estoppel?

157. In the light of my conclusions about Grattan (ECJ), I can be fairly brief in my consideration of HMRC's first overarching contention. It will be recalled that I do not consider Grattan (ECJ) to have had anything like the radical impact on the underlying liability issue that HMRC would have the court believe. In my view its significance is relatively marginal, and although the analysis by Advocate General Kokott of the wider background issues may suggest a number of doubts about aspects of the decision in Littlewoods (CA), her analysis was not endorsed by the ECJ except in relation to a point which has never been in dispute. Furthermore, her analysis did not depend to any appreciable extent on EU case law which post-dated Littlewoods (CA); and the one point on which she did apparently differ from the Court of Appeal's reasoning was in my view of little significance.
158. In truth, it seems to me that the real relevance of Grattan (ECJ) is that it belatedly suggested to HMRC, and gave some colour to, a number of arguments on liability which in principle could perfectly well have been raised as grounds for seeking leave to appeal to the House of Lords from the decision in Littlewoods (CA), and (in my view at least) would have offered some prospects of success had leave been granted. What cannot be said with any plausibility, in my judgment, is that Grattan (ECJ) showed for the first time that Littlewoods (CA) had been wrongly decided, or that the relevant legal principles had changed in any major respects, or (as Mr Swift put it in his opening submissions) that Grattan (ECJ) represented a sea change in the way in which the relevant commission arrangements were to be analysed.
159. One of the points which Mr Swift made in this connection was that Grattan (ECJ) provided the first occasion for the ECJ to consider commission arrangements which were either the same as, or very similar to, those in the present case. In my view this point carries little weight, for two reasons. First, the ECJ had already considered arrangements of a very similar nature in Freemans, albeit in the context of a dispute about timing. Secondly, the only issue which the Court actually had to determine in Grattan (ECJ) was a narrow one, confined to the period between 1973 and 1978.
160. Assuming for present purposes that the conditions for issue estoppel to apply are all satisfied, the question is whether the decision in Grattan (ECJ) is of sufficient significance to bring the case within the exception which the House of Lords recognised in Arnold. As Lord Sumption pointed out in Virgin Atlantic at [20], the real issue in Arnold was whether the flexibility in the doctrine of res judicata, which was implicit in Wigram V.-C.'s statement of the law in Henderson v Henderson 3 Hare 100 at 115, extended to an attempt to re-open the very same point in materially altered circumstances, as well as to cases where the point in question had not been considered and decided, but arguably should have been.
161. It was in that context that Lord Keith formulated the exception at 108G-109C:

"But there is room for the view that the underlying principles upon which estoppel is based, public policy and justice, have greater force in cause of action estoppel, the subject matter of the two proceedings being identical, than they do in issue estoppel, where the subject matter is different. Once it is accepted that different considerations apply to issue estoppel, it is hard to perceive any logical distinction between a point which was previously raised and decided and one which might have been but was not. Given that the further material which would have put an entirely different complexion on the point was at the earlier stage unknown to the party and could not by reasonable diligence have been discovered by him, it is hard to see why there should be a different result according to whether he decided not to take the point, thinking it hopeless, or argue it faintly without any real hope of success. In my opinion your Lordships should affirm it to be the law that there may be an exception to issue estoppel in the special circumstance that there has become available to a party further material relevant to the correct determination of a point involved in the earlier proceedings, whether or not that point was specifically raised and decided, being material which could not by reasonable diligence have been adduced in those proceedings. One of the purposes of estoppel being to work justice between the parties, it is open to courts to recognise that in special circumstances inflexible application of it may have the opposite result, as was observed by Lord Upjohn in the passage which I have quoted above from his speech in the Carl Zeiss case [1967] 1 AC 853-957."

162. In his concurring speech, Lord Lowry said at 112G that it was only "in exceptional circumstances" that the general rule should be relaxed. The need for caution in recognising an exception to the rule was also emphasised by Lord Neuberger of Abbotsbury PSC in Virgin Atlantic at [62]:

"When seeking to justify a conclusion that, though it applies, res judicata does not preclude a point being taken, it can be dangerous to invoke the observation of Lord Keith of Kinkel in Arnold v National Westminster Bank Plc [1991] 2 AC 93, 109B, that estoppel is intended "to work justice between the parties", because it is only too easy to fall back on it as an excuse for an unprincipled departure from, or an unprincipled exception to, the rule."

163. Furthermore, as Lord Hoffmann pithily observed in Carter v Ahsan [2007] UKHL 51, [2008] 1 AC 696, at [33], in a speech with which the other four members of the court agreed:

"The whole point of an issue estoppel on a question of law is that the parties remain bound by an erroneous decision."

If the Arnold exception is to operate, it is clearly not enough that the previous decision on an issue of law has been shown by a subsequent decision to be erroneous. The new material relied upon must, in addition, be "material which could not by reasonable diligence have been adduced in [the earlier] proceedings": Arnold at 109B. Where the new material relied upon is a change in the law, the right approach was indicated by Lord Keith at 109F-G:

"If a judge has made a mistake, perhaps a very egregious mistake, as is said of Walton J's judgment here, and a later judgment of a higher court overrules his decision in another case, do considerations of justice require that the party who suffered from the mistake should be shut out, when the same issue arises in later proceedings with a different subject matter, from re-opening that issue?"

164. The facts in Arnold were striking. Under a sub-underlease for a term of 32 years of commercial premises entered into in 1976, there was provision for rent reviews at approximately five-yearly intervals, the amount of the reviewed rent to be fixed by reference to a hypothetical lease for the unexpired residue of the term. On the first review, in 1983, there was a dispute whether the hypothetical lease was to be construed as itself containing the same rent review provisions as the actual lease, or no provision for review. Walton J held, on an appeal from an arbitrator, that it was to be construed as containing no provision for review. He refused the tenant permission to appeal, and also refused to grant a certificate under section 1(7)(b) of the Arbitration Act 1979 that there was a question of law of general public importance or one which for some other special reason ought to be considered by the Court of Appeal. The Court of Appeal subsequently held that it had no jurisdiction to entertain an appeal from the refusal to grant a certificate.
165. At that point, therefore, the tenant had exhausted the legal remedies available to it. Later judicial decisions, however, including two in the Court of Appeal, showed that Walton J had almost certainly erred in his construction of the relevant clause in the lease, and when the second rent review fell due in June 1988, the question arose whether the tenant was barred by issue estoppel from re-litigating the point decided by Walton J. The question was decided in the tenant's favour by Sir Nicolas Browne-Wilkinson V.-C., the Court of Appeal, and the House of Lords.
166. Among the factors which weighed with Lord Keith in reaching this conclusion were:

(a) the fact that there was no right of appeal against Walton J's judgment;

(b) the injustice to the tenant of being faced with a succession of rent reviews over a period of more than 20 years, all proceeding on a construction of the lease which was highly unfavourable to the tenant and was generally regarded as wrong;

(c) the fact that the limitations on appeals introduced by the Arbitration Act1979 were not present at the commencement of the lease in 1976; and

(d) little weight could be attributed to the public interest in seeing an end to litigation, given that there anyway had to be an arbitration at each successive review date.

Taking these factors into account, together with the unmerited windfall which the landlord would enjoy if the issue could not be re-opened, Lord Keith concluded at 110G:

"Estoppel per rem judicatam, whether cause of action estoppel or issue estoppel, is essentially concerned with preventing abuse of process. In the present case I consider that abuse of process would be favoured rather than prevented by refusing the plaintiffs permission to re-open the disputed issue."

167. Mr Rabinowitz referred me to subsequent cases of high authority in which a strict approach has been taken to recognition of cases which might fall within the Arnold exception. One of those cases was Carter v Ahsan: see the speech of Lord Hoffmann at [33] to [35]. Another was the decision of the Court of Appeal in Curling and Others v Securicor Ltd and Another [2001] EWCA Civ 358, apparently unreported, where Peter Gibson LJ (with whom Mummery and Rix LJJ agreed) at [34] described Arnold as "a case on its own very special facts", and at [35] distinguished the facts of the instant case as being "not nearly so compelling". He added at [36]:

"In Arnold at first instance the Vice-Chancellor said ([1989] Ch at page 70) that there was great danger in allowing allegations that the law had changed. I respectfully agree. In my judgment, even if I had thought that there was a change in the law, I would have held that issue estoppel should apply."

168. Applying these principles, and in the light of my views about the relative insignificance of Grattan (ECJ), I am left in no real doubt that HMRC's first overarching contention must be rejected. The circumstances of the present case do not in my opinion come within measurable distance of the kind of change in the law which may, exceptionally, prevent an issue estoppel from operating. Although not by itself conclusive, the fact that HMRC did not seek leave to appeal from the decision in Littlewoods (CA) tells strongly against permitting the issues decided in that case to be re-opened. So too does the fact that HMRC freely agreed to enter into the 2004 and 2008 section 85 Agreements. I accept that the balance might arguably have come down in HMRC's favour if Grattan (ECJ) were indeed a ground-breaking decision which clearly demonstrated for the first time that the tax in question was due; but that is very far from being the position.

(c) The application of issue estoppel to tax cases

169. In considering the Revenue's second overarching contention, it is convenient to begin with the judgment of the Privy Council in Caffoor, on appeal from the Supreme Court of Ceylon. The underlying issue in the case was whether the income of a trust established in Ceylon in 1942, the Abdul Gaffoor Trust, was exempt from liability to income tax, after the death of the grantor in 1948, on the ground that the trust was an "institution or trust of a public character established solely for charitable purposes" within the meaning of section 7(1)(c) of the Ceylon Income Tax Ordinance of 1932. Under the relevant trusts, the trustees had a wide discretion to apply the trust income for a number of purposes which, viewed in isolation, were of a broadly charitable nature, but in so doing they were required to give priority to male descendants of the grantor.
170. The trustees were assessed to income tax for the revenue year 1949/50 on the footing that the exemption did not apply, but they appealed successfully to the Board of Review which gave its decision in December 1954, holding that the income fell within the statutory exemption. Precisely the same issue then arose on appeals from assessments on the trustees for the five succeeding revenue years (1951/52 to 1954/55), and the trustees argued that it had already been conclusively determined in their favour by the previous decision of the Board of Review. This argument was rejected at all levels in the courts of Ceylon, but for differing reasons. As Lord Radcliffe, delivering the judgment of the Privy Council, explained at p 597, one ground for rejecting the estoppel had been that the Board of Review was an essentially administrative body which did not perform judicial functions, whereas another ground had been that a decision on an assessment for a previous year was incapable of binding the Board of Review in a subsequent year. Different consequences might, of course, follow, depending on which view was correct. If the former, a decision of the Supreme Court, which indubitably exercised judicial functions, would be capable of grounding an issue estoppel; but if the latter view were correct, "a Supreme Court decision would be no more capable of setting up an estoppel than would one made by the Board of Review, whatever its precise status as a judicial tribunal".
171. Lord Radcliffe continued:

"In their Lordships' opinion the question of estoppel cannot be decided merely by enquiring to what extent the Board of Review exercises judicial functions. The critical test is not the bare issue whether or not such a board exercises judicial power … What is important here is that the Board of Review is a tribunal set up under the Income Tax Ordinance for the purpose of deciding income tax appeals at a certain stage of their prosecution, and that decisions given with regard to such appeals are effective only within the limited jurisdiction that the Ordinance creates for all tribunals that deal with the matter of an appeal. All such appeals remain in one sense a part of the process of assessment since all the tribunals, including the Supreme Court, have independent power to increase or reduce the assessment under appeal. While, therefore, it is unexceptionable to say that the Board of Review when exercising its powers under section 73 is acting in a sense judicially, that the dispute which it has to determine is at any rate somewhat analogous to a lis inter partes and that the assessor who made the assessment or some other representative of the Commissioner … resembles a party hostile to the appellant, these considerations are not those that are critical to the issue of estoppel. The critical thing is that the dispute which alone can be determined by any decision given in the course of these proceedings is limited to one subject only, the amount of the assessable income for the year in which the assessment is challenged. It is only the amount of that assessable income that is concluded by an assessment or by a decision on an appeal against it (see section 75). Although, of course, the process of arriving at the necessary decision is likely to involve the consideration of questions of law, turning upon the construction of the Ordinance or of other statutes or upon the general law, and the tribunal will have to form its view on those questions, all these questions have to be treated as collateral or incidental to what is the only issue that is truly submitted to determination …"

172. Lord Radcliffe drew support for this analysis from a rating case decided by the Privy Council in 1926, Broken Hill Proprietary Co Ltd v Broken Hill Municipal Council [1926] AC 94, where a valuation and liability to tax in one year was treated as a different question from the valuation and liability to tax in a subsequent year, with the result that the principle of res judicata could not apply. The case was a striking one, because the earlier judgment which was said to ground the estoppel had been given by the High Court of Australia. As Lord Radcliffe observed at p 599:

"It underlines the point that it is not the status of the tribunal itself, judicial or administrative, that forms the determining element for estoppel in cases of this kind but the limited nature of the question that is within the tribunal's jurisdiction. The judgment of the High Court that had been given in the earlier year was explicitly directed to the construction of a particular section of the rating Act and to the correct measurement of the liability in the light of that construction. Precisely the same point arose in the later year and was ultimately decided by this Board in a sense contrary to that which had previously been adopted."

173. Lord Radcliffe went on to explain that, in all the cases which had arisen under income tax or rating appeal procedure, the decision was "essentially as to the correct amount (if any) of the assessment", and to apply the Broken Hill principle to the Caffoor case would be "to bring it into line with what seems to be by now the regular course of authority with regard to appeals in successive years against income tax or rating assessments". He continued:

"It may be that the principles applied in these cases form a somewhat anomalous branch of the general law of estoppel per rem judicatam and are not easily derived from or transferred to other branches of litigation in which such estoppels have to be considered; but in their Lordships' opinion they are well established in their own field, and it is not by any means to be assumed that the result is one that should be regretted in the public interest."

174. Despite Lord Radcliffe's reference to the public interest, it is I think clear that the decision turned on the correct identification of the issue which the Board of Review had jurisdiction to determine. If that issue was, as a matter of construction of the relevant legislation, confined to determination of the amount of assessable income for the relevant year of assessment, it could not, by definition, give rise to an estoppel in relation to any other year. So viewed, it may be said (and Mr Rabinowitz duly submitted) that the doctrine is not truly anomalous, because it merely illustrates the general principle that no issue estoppel can arise unless the same issue has been determined on an earlier occasion. On the other hand, it is not so easy to explain why an estoppel is still incapable of arising even where, as part of its determination for the earlier year, the court or tribunal has decided a question, whether of fact or law, which is in all material respects identical to one which arises in the subsequent year. To say, as Lord Radcliffe did, that all such questions have to be treated as "collateral or incidental" to the only issue truly submitted to determination, is in substance to introduce a further requirement which has no parallel in the general law of issue estoppel. In my view, a further requirement of this nature may indeed be regarded as anomalous, and (assuming it to be sound in principle) it invites the question whether it may have a different justification.
175. Anomalous or not, there is in my judgment no doubt that the Caffoor principle remains good law in England and Wales, at least in relation to income tax, corporation tax, capital gains tax and other annually assessed (or, nowadays, self-assessed) taxes, where the basic question for determination is the correct amount of tax payable for the relevant year or period of assessment. The principle has been consistently applied at first instance by the High Court in a number of cases: see, for example, Tod v South Essex Motors [1988] STC 392 at 408j-409c (Knox J), Barnett v Brabyn (Inspector of Taxes) [1996] STC 716 at 723c-j (Lightman J), MacNiven v Westmoreland Investments Ltd [1997] STC 1103 at 1133 (Carnwath J) and King v Walden (Inspector of Taxes) [2001] STC 822 (Jacob J, discussed in more detail below).
176. In the Westmoreland case, the issue was whether an agreement which the taxpayer company had entered into with the Inspector of Taxes under section 54 of the Taxes Management Act 1970 ("TMA 1970") (the equivalent provision for most direct taxes of section 85 of VATA 1994) for the year ended 31 March 1988 determined not only how much tax was payable in respect of the period covered by the assessment under appeal, but also the amount of charges on income which were available to be carried forward to subsequent accounting periods. The main issue in the case, irrelevant for present purposes, concerned the application of the Ramsay principle to a series of circular transactions designed to transform an unpaid liability for interest into deductible charges on income. The Court of Appeal held, reversing Carnwath J, that the scheme succeeded in its object, with the result that it was unnecessary to deal with the section 54 agreement point. Nevertheless, Peter Gibson LJ said, having heard full argument on the point, that in his judgment the judge was plainly right in rejecting the taxpayer's argument for the reasons he had given: see [1998] STC 1131 at 1147a-b. In the House of Lords, the decision of the Court of Appeal on the main issue was upheld, but Lord Hope of Craighead dealt obiter with the section 54 issue, reaching the same conclusion as both courts below: see MacNiven v Westmoreland Investments Ltd [2001] UKHL 6, [2003] 1 AC 311, at [82] to [90]. The only other member of the court to address the issue was Lord Hoffmann, who at [75] expressed his agreement with Lord Hope's reasoning.
177. After explaining that the effect of a section 54 agreement was to attach the same finality to the settling of appeals by agreement as to the determination of the appeal by the general or special commissioners, and quoting from Lord Radcliffe's judgment in Caffoor at p 598, Lord Hope said this:

"89. I would apply that reasoning to the present case. The Ceylon Ordinance was closely modelled on the legislation that applies in the United Kingdom. The purpose of an appeal under section 31 of [TMA 1970] is to challenge the amount charged to tax by an assessment. The finality that attaches to the determination of the appeal by the general commissioners or by the special commissioners or to the settling of the appeal by agreement relates only to the amount chargeable under that assessment. The question as to the amount of any reliefs carried forward to subsequent periods remains open for examination as the assessment for each subsequent period is issued. This is because the Taxes Acts do not provide any means by which that amount may be determined conclusively, whether by appeal or by agreement, for any period other than that to which the assessment relates.

90. For these reasons I would hold that an agreement made under section 54 has no wider effect upon the position of either party than that which has been provided for by the statute. As Carnwath J … indicated, the issue turns simply and solely upon the machinery which the Taxes Acts provide for determining the amount in question between the commissioners and the taxpayer. That machinery is limited to determining conclusively the amount of tax chargeable for the year of assessment. It does not enable such determinations to be made, either on appeal or by agreement, as to the amounts of tax chargeable in future years."

178. The decision of Jacob J (as he then was) in King v Walden, loc. cit., provides a graphic illustration of the reach of the Caffoor principle. In 1991 the special commissioners upheld on appeal various "out of time" assessments to income tax which had been made on the taxpayer, Mr King, finding that he had been guilty of wilful default or neglect in respect of the relevant years. In due course, associated determinations of penalties and interest were made against him, and in 1996 a further set of out of time assessments was made on him following the discovery of a previously undisclosed asset. Mr King's appeals against these further determinations and assessments were eventually heard, and substantially dismissed, in 2000 by a differently constituted panel of special commissioners. In dealing with the question of interest, they held that they were bound by the findings of wilful default and neglect which had been made by their predecessors in 1991, and that they were also bound (in relation to both the in time and the out of time assessments) by the earlier determinations of the amounts of tax lost.
179. On further appeal to the High Court, it was argued for Mr King that the commissioners in 2000 were wrong in law about the binding effect of the 1991 decision, and that it was open to him to reargue not only the question of wilful default or neglect for the relevant years, but also (in relation to the determinations of interest) the amounts of tax held due in 1991 on which the interest ran. The argument was that the only thing which Mr King could not dispute was the determinations of tax due which the commissioners had actually made in 1991.
180. Jacob J evidently found this a startling proposition. As he said, at [15] of his judgment:

"It means that one could have a first finding that tax was due yet a subsequent finding that no interest or that any tax was due because tax had not been due after all. Or, as is claimed here, a finding of wilful default or neglect justifying out-of-year assessments but then a finding of no interest on those assessments because wilful default or neglect was not re-proved. Moreover the Revenue, or indeed the taxpayer, would find itself having to prove the same thing over and over again in relation to exactly the same facts for exactly the same periods."

181. Notwithstanding his initial surprise, however, Jacob J felt constrained by authority to uphold Mr King's argument. Having reviewed the case law, including Caffoor and Westmoreland, and having noted the statements by Lord Radcliffe and Lord Hope in those cases to the effect that it is only the amount of assessable income that is concluded by an appeal (see [19]), he stated his conclusion as follows:

"27. But the opinion in Caffoor has been taken as representing the law in many cases by now. I have enormous sympathy with the view that once a matter is decided after a full and fair fight that is that. I can see no real reason for a different rule for tax cases. But I think I must, as a judge of first instance, bow to the weight of authority which does not distinguish between settled and fought appeals. Accordingly I hold that the 2000 commissioners were wrong to apply s46(2) [of TMA 1970, which says that the determination of the general or special commissioners in any proceedings under the Taxes Acts shall be "final and conclusive"] to the interest appeals.

28. There is one other point worth mentioning in relation to the effect of s46(2) which indirectly supports its non-binding effect. Suppose an original assessment upheld by commissioners. The taxpayer can only appeal to the court on a point of law. Then suppose a penalty determination based on those assessment again upheld by commissioners. From a penalty determination appeal lies to the court on fact and law. But if s46(2) (or the general law of issue estoppel) excluded any challenge to the earlier assessment and particularly to any point necessarily determined to uphold that earlier assessment (e.g. wilful default) then the right of appeal to the court on fact in respect of penalties would be nugatory or emasculated. Mr Woolf [counsel for Mr King] characterised this result as absurd."

182. Mr Swift relies heavily on King v Walden, arguing that it provides a close and cogent parallel to the present case. Here too, he says, the issues of liability determined (or deemed to have been determined) in Littlewoods' favour related only to the original payments of VAT. Entitlement to interest (whether statutory, or by way of a restitutionary claim) is a different matter, and (as in King v Walden) it should be open to the Revenue to defend the separate claims for interest by arguing that the relevant VAT had always been due and should never have been repaid.
183. This submission therefore raises the question whether the Caffoor principle (both in general, and as applied to interest in King v Walden) is part of the law of VAT in the same, or an analogous, way as it is part of the law of income tax. If the answer to that question is affirmative, Mr Swift's argument seems to me a very forceful one. If the answer is negative, however, any justification for departure from the general law of issue estoppel must be found elsewhere.
184. In structural terms, there are some obvious differences between income tax (and other assessed, or self-assessed, direct taxes) on the one hand, and VAT on the other hand. The former are assessed on an annual basis, in accordance with the familiar machinery laid down in TMA 1970 and related statutes. The essentially periodic nature of the process is reinforced by the fact that income tax and corporation tax (although not capital gains tax or inheritance tax) are annual taxes imposed by Parliament each year: see Tiley and Loutzenhiser, Revenue Law, 7^th ed (2012) at para 2.1.3. On an appeal from an assessment (or an amendment to a self-assessment), the ultimate issue is normally the correct amount of tax for the year of assessment in question: see TMA 1970 section 50, the basic provisions of which have remained essentially unchanged over the years, although since 2009 tax appeals have been heard by the FTT instead of the general or special commissioners. In reaching its conclusion, the tribunal or court hearing the appeal is free to form its own view on the law, unconstrained (subject to the requirements of procedural fairness and proper case management) by the arguments of the parties, the history of the dispute, or the amount of tax assessed. This last point reflects the important general public interest in taxpayers paying the correct amount of tax, on which see the observations of Lord Walker and Lord Hope in Tower MCashback LLP 1 v Revenue and Customs Commissioners [2011] UKSC 19, [2011] 2 AC 457, at [15] to [18] and [84] to [85] respectively.
185. VAT, by contrast, is a permanent tax imposed by EU legislation, which is reflected in, and implemented by, domestic legislation in the UK (both primary and secondary). For most of the period covered by the claims, the principal EU legislation was contained in the First and Sixth VAT Directives (both of which were superseded with effect from 1 January 2007 by Council Directive 2006/112/EC on the Common System of Value Added Tax). VATA 1994 (as amended) is still the principal source of domestic primary legislation. For present purposes, it is enough to refer to the provisions of the Sixth Directive and VATA 1994.
186. VAT is a tax on "the supply of goods or services effected for consideration within the territory of the country by a taxable person acting as such": article 2(1) of the Sixth Directive. It becomes chargeable when the goods are delivered or the services are performed (article 10(1)), and the right to deduct input tax arises at the same time (article 17(1)). The tax is collected and accounted for in the UK by the familiar system of periodic returns and payments prescribed by Schedule 11 to VATA 1994 and Part V of the VAT Regulations 1995 (SI 1995 No. 2518). Article 22(4) of the Sixth Directive permits Member States to determine the tax periods for which returns must be submitted, the basic options being one month, two months or a quarter, with an upper limit of one year.
187. Section 80 of VATA 1994 deals with the recovery of overpaid VAT. The version of the section in force before 20 July 2005 provided, by subsection (1), that:

"Where a person has (whether before or after the commencement of this Act) paid an amount to the Commissioners by way of VAT which was not VAT due to them, they shall be liable to repay the amount to him."

The version of the section currently in force (from 20 July 2005 to date) replaces the right to a repayment with a right to be credited the amount overpaid, and it applies where a person:

"(a) has accounted to the Commissioners for VAT for a prescribed accounting period (whenever ended), and

(b) in doing so, has brought into account as output tax an amount that was not output tax due."

188. The former version of section 80(1) did not refer expressly to payments of VAT made for prescribed accounting periods, but I do not attribute any significance to this difference. Any VAT which has been overpaid will inevitably have been paid in the context of the system of periodic returns and accounting which has applied ever since the tax was first introduced in 1973. The express reference to accounting periods in the new version of the section is in my view attributable to the introduction of a right to a credit instead of a right to immediate repayment. Administration of a credit system needs to be tied in with the system of periodic accounting.
189. Sections 73 and following of VATA 1994 empower HMRC to make assessments of the amount of VAT due in a variety of circumstances, including failure to make returns, or where it appears to HMRC that returns are incomplete or incorrect. Provision is also made for the imposition of interest, penalties and surcharges, and (in section 76) for the assessment of amounts due in respect thereof. Appeals are dealt with by section 83, subsection (1) of which provides that an appeal shall lie to the Tribunal with respect to a long list of certified matters, including:

"(b) the VAT chargeable on the supply of any goods or services;

…

(p) an assessment … or the amount of such an assessment;

(q) the amount of any penalty, interest or surcharge specified in an assessment under section 76;

…

(s) any liability of the Commissioners to pay interest under section 78 or the amount of interest so payable;

…

(t) a claim for the crediting or repayment of an amount under section 80;

…"

Section 84 then contains further provisions relating to the different categories of appeal. There is no general equivalent of section 50 of TMA 1970, but in relation to appeals on penalties etc it is expressly provided that nothing in the relevant part of section 83 shall be taken to confer on a tribunal any power to vary an amount assessed by way of penalty etc "except in so far as it is necessary to reduce it to the amount which is appropriate…": see subsection (6).

190. In the light of the statutory scheme of VAT which I have outlined, I can see no good reason why the Caffoor principle, with suitable modifications, should not apply to it in a similar way, at least where the dispute relates to the amount of VAT chargeable on supplies of goods or services in one or more (usually quarterly) periods, or to assessments (whether of VAT, interest, penalties or surcharges) made for particular periods, or to claims for the repayment of VAT originally paid in respect of particular periods. In all these cases, the periodic framework in which the issue arises may be reasonably regarded as analogous with the yearly assessment of income tax and other direct taxes, although there is of course no equivalent to years of assessment as such, and no limit to the number of VAT accounting periods which may be aggregated in a single assessment or determination. To my mind, however, these are differences of machinery rather than substance, and the more significant points are that VAT is in essence a transaction-based tax which is returned and accounted for on a periodic basis.
191. This conclusion is in line with two decisions of the Tribunal to which Mr Swift referred me: SITA, [2002] UKVAT V17991, [2003] V & D R 131, and Durwin Banks, [2008] UKVAT V20695, released on 29 May 2008. The issue in both cases was whether a particular supply of goods or services (telecommunication services to the aviation industry in SITA, unrefined linseed oil in Durwin Banks) should be zero-rated. In SITA, the Tribunal, chaired by Mr Stephen Oliver QC, concluded at [70] that "issue estoppel has no place in VAT litigation of this nature". The Tribunal was influenced by the principle of public policy that the tax should operate uniformly ([68]), as well as by the fact that the earlier decision in favour of zero-rating had been made by consent 30 years earlier, without argument or reasoning, when VAT in the UK was still in its infancy. In Durwin Banks, the earlier decision had been more recent (2005), but the appellant had been unrepresented at the hearing. The Tribunal, chaired by Mr Theodore Wallace, reviewed the Caffoor principle and reached the same conclusion as in SITA. It was influenced, rightly in my view, by the EU law principle of fiscal neutrality, pointing out at [48]:

"Since we have concluded that another trader not fettered by res judicata would succeed on the material before us in establishing that similar bottled linseed oil is food, a decision against this Appellant based on res judicata would conflict with the principle of fiscal neutrality."

The Tribunal also observed that repeated attempts to re-litigate the same issue without good reason could be controlled by the principle of abuse of process, which was much less inflexible than issue estoppel. Again, I respectfully agree.

192. Since VAT is governed by EU law, it is obviously material to enquire whether the ECJ has given any relevant guidance on the scope for issue estoppel in relation to VAT. The answer is that it has, in Case C-2/08, Amministrazione dell'Economia e delle Finanze and Agenzia delle Entrate v Fallimento Olimpiclub Srl [2009] ECR I-7501 ("Olimpiclub"), where the Second Chamber of the ECJ had to consider the impact of the principle of res judicata in national law on questions relating to VAT. Article 2909 of the Italian Civil Code, entitled "Res judicata", provides that: "Findings made in judgments which have acquired the force of res judicata shall be binding in all respects on the parties, their lawful successors or assignees". As interpreted by the Italian Supreme Court of Cassation in 2006, article 2909 applied wherever an earlier set of proceedings between the same parties had given rise to a judgment resolving points of either fact or law on a fundamental issue, common to both cases, even if the aims of the later proceedings were different from those of the first.
193. The facts of the case were briefly as follows. The taxpayer company, Olimpiclub, owned a sports complex on land belonging to the Italian State. In 1985 it entered into a contract with a closely related non-profit making entity ("the Association") under which the Association was granted the use of all the facilities of the sports complex, in return for which it agreed to transfer its entire gross income to Olimpiclub. The fees paid by the club members to the Association were not liable to VAT, and in this way it was hoped to avoid the VAT which would have been charged if the relevant services had been provided directly by Olimpiclub to the members. The Italian tax authorities took the view that "the parties to the contract had, in reality, by means of an act which on the face of it was lawful, intended solely to circumvent the legislation in order to obtain a tax advantage" (paragraph 6 of the judgment of the ECJ), and made adjustments accordingly to the VAT returns submitted by Olimpiclub for the tax years 1988 to 1991. Olimpiclub challenged the adjustment notices before the Provincial Tax Court of First Instance, which found in Olimpiclub's favour on the basis that the agreement had not been shown to be fraudulent. The Finance Administration appealed to the Regional Tax Court, but the appeal was dismissed. On a further appeal to the Court of Cassation, Olimpiclub (now in liquidation) sought to rely on two judgments of the Regional Tax Court which arose out of the same investigation but related to different tax years (1992 and 1987 respectively), and which had acquired the force of res judicata pursuant to article 2909 of the Civil Code.
194. It appears from paragraphs 12 to 18 of the judgment of the ECJ that, when interpreting article 2909 in the context of taxation matters, the Italian courts had for a long time adhered to a principle very similar to the Caffoor principle, but this approach had recently been modified to allow scope for the principle of res judicata in relation to a different tax period. The Court of Cassation regarded itself as bound by the earlier judgments which had held the agreement in question to be genuine, lawful and not fraudulent; but it was concerned that this could preclude it from examining the question in light of the doctrine of abuse of rights as recognised in the Halifax case (Case C-255/02, Halifax Plc and Others, [2006] ECR I-1609, [2006] Ch 387). As the ECJ commented:

"17. Since the levying of VAT plays a central role in the accrual of the European Community's own resources, the referring court is uncertain whether the case-law of the Court requires the binding authority of a judgment which has acquired the force of res judicata by reason of national law to be disregarded. In the main proceedings, the application of Article 2909 of the Italian Civil Code could prevent the full implementation of the principle of combating abuse of rights, a principle which the Court has developed in relation to VAT as a means of ensuring that the Community VAT system is fully implemented."

195. As reformulated by the ECJ in paragraph 19, the question referred to it by the Court of Cassation asked:

"… in essence, whether Community law precludes the application, in circumstances such as those at issue in the main proceedings, of a provision of national law, such as Article 2909 of the Italian Civil Code, in a VAT dispute in relation to a tax year for which no final judicial decision has yet been delivered, where that provision would prevent the referring court from taking into consideration Community law rules concerning abusive practice in the field of VAT."

In considering this question, the Court began by recognising the importance, both for the EU legal order and for national legal systems, of the principle of res judicata, and confirmed that EU law does not require a national court to disapply domestic rules of procedure conferring finality on a decision, even if to do so would make it possible to remedy an infringement of EU law: see paragraphs 22 and 23. The Court then restated its familiar jurisprudence, to the effect that in the absence of relevant EU legislation, such procedural matters were for Member States to determine in accordance with the principle of procedural autonomy, provided that the rules in question complied with the EU principles of equivalence (i.e. they must not be less favourable than those governing similar domestic actions) and effectiveness (i.e. they may not be framed in such a way as to make it in practice impossible or excessively difficult to exercise the rights conferred by EU law).

196. The Court then considered whether the recognition of res judicata in VAT matters was compatible with the principle of effectiveness, bearing in mind the importance of the principle of legal certainty in both national and EU law. The nub of the Court's decision is contained in the following paragraphs:

"28. It is therefore necessary to determine, specifically, whether the above mentioned interpretation of Article 2909 of the Italian Civil Code may be justified with a view to protecting the principle of legal certainty, in the light of its implications for the application of Community law.

29. It should be noted that - as the national court itself points out – not only does the interpretation in question prevent a judicial decision that has acquired the force of res judicata from being called into question, even if that decision entails a breach of Community law; it also prevents any finding on a fundamental issue common to other cases, contained in a judicial decision which has acquired the force of res judicata, from being called into question in the context of judicial scrutiny of another decision taken by the relevant tax authority in respect of the same taxpayer or taxable person, but relating to a different tax year.

30. Accordingly, if the principle of res judicata were to be applied in that manner, the effect would be that, if ever the judicial decision that had become final were based on an interpretation of the Community rules concerning abusive practice in the field of VAT which was at odds with Community law, those rules would continue to be misapplied for each new tax year, without it being possible to rectify the interpretation.

31. In those circumstances, it must be held that such extensive obstacles to the effective application of the Community rules on VAT cannot reasonably be regarded as justified in the interests of legal certainty and must therefore be considered to be contrary to the principle of effectiveness."

197. The Court has therefore held, in terms, that res judicata (in the sense of issue estoppel) cannot prevent application of the EU doctrine of abuse of rights in the field of VAT on a year by year basis. In such circumstances, the principle of legal certainty is trumped by the principle of effectiveness. The Court did not decide, no doubt because it did not need to, whether the same reasoning would apply in relation to other alleged breaches of EU law, but it must in my view be strongly arguable that the same approach should be adopted whenever an issue of liability to VAT arises. I cannot see any satisfactory reason for distinguishing between cases where the national tax authority wishes to rely on the doctrine of abuse of rights, and cases where it wishes to argue that, as a matter of EU law, an earlier decision in the taxpayer's favour on an issue of liability is mistaken. Different considerations may arise in relation to issues of fact, but I need not investigate them further because the estoppels on which Littlewoods wish to rely are all estoppels of law arising from agreed facts.
198. I have yet to mention one further case on which Mr Swift placed considerable reliance, namely the decision of Christopher Clarke J (as he then was) in a case about the payment of customs duty on imported swimwear, Matalan Retail Ltd v Revenue and Customs Commissioners [2009] EWHC 2046 (Ch), [2009] STC 2638. The facts were complicated, and involved issues of abuse of process in the broader sense, as well as potential issue estoppels in relation to the correct tariff applicable to the swimwear based on a binding tariff information ("BTI") which the Revenue had issued to the taxpayer in 2006 and from which it subsequently sought to resile. The judge decided the broader abuse of process arguments in the Revenue's favour, and also decided that there had never been any judicial determination of the validity or otherwise of the "essential proposition" that swimwear with no rubber in the synthetic fibre content, but with rubber amounting to 5% or more of the garment as a whole, was properly to be given a 10 and not a 90 classification under the relevant part of the Combined Nomenclature set out in the annually revised Annex 1 to EC Council Regulation 2658/87. The judge then held that, even if the Tribunal had decided the essential proposition, neither it nor the Revenue would have been bound by that determination, as a matter of res judicata, in relation to future transactions or BTIs. It is this part of the judgment, strictly speaking obiter, which is relevant for present purposes.
199. After a careful review of the Caffoor line of authority, and the decision of the House of Lords in Vandervell Trustees Ltd v White [1971] AC 912, (1970) 46 TC 341, the judge summarised the conclusions which he drew from the English case law:

"116. The upshot of these decisions would appear to be as follows:

(a) a decision by a tribunal charged with determining the correct rate or tax to be paid in respect of one year is not to be treated as determining the rate or tax to be paid in a succeeding year even if there is no change in circumstances, whether of fact or law. By parity of reasoning a decision by a tribunal as to the duty to be paid on a particular importation of goods is not determinative of the duty to be paid on a different importation of goods even if there is no change of circumstances as between the two importations. The tribunal and the parties concerned (tax authority and taxpayer) are not bound by any estoppel created by the previous decision either as to law or fact;

(b) if the tax authorities and the taxpayer are parties to a court case in which the court has determined a question of fact, the parties are estopped from putting forward evidence or making submissions which are inconsistent with the determination of that question by the court. But a tribunal whose task it is to determine the tax, rate or duty payable is not so bound.

117. I doubt whether proposition (a) can continue to be justified by any notion that the tribunal is not the sort of judicial body whose decisions could constitute res judicata … Proposition (b) would appear to follow from Vandervell and ordinary principles … "

200. In [118] and [119], the judge referred to, and quoted from, the decision of the US Supreme Court in US v Stone & Downer & Co 274 US 225 (1927), where it was held that the principles of res judicata had only restricted application to judicial decisions on customs duty. If it were otherwise, said the US Supreme Court:

"[t]he importing house which has … obtained the favourable position permanently binding on the Government will be able to import the goods at a much better rate than that enjoyed by … its competitors. Such a result would lead to inequality in the administration of the customs law, to discrimination and to great injustice and confusion. In the same way if the first decision were against a large importing house, and its competitors … succeeded in securing a different conclusion, the first litigant, bound by the judgment against it … must permanently do business … at great and inequitable disadvantage with its competitors."

In [121], Christopher Clarke J expressly recognised that "it is in the public interest that traders should pay the correct amount of tax or duty – no more and no less".

201. In answer to these submissions, Littlewoods argue that the Caffoor principle is anomalous and should not be extended beyond the context of recurring annual or periodic taxes. Mr Rabinowitz referred me to the comments of Laskin J in a dissenting judgment in the Supreme Court of Canada in 1974, when he doubted whether the principle "retains any survival value" and said he "would reject the introduction of such an anomaly into the law of Canada": see Angle v Minister of National Revenue [1975] 2 SCR 248 at 266. Mr Rabinowitz also submitted that, if the scope of the principle were extended to VAT, this would be inconsistent with the decision of the House of Lords in Thrasyvoulou v Secretary of State for the Environment [1990] 2 AC 273, where Lord Bridge of Harwich said at 289C-D that the principles of res judicata "must apply equally to adjudications in the field of public law". He then said:

"In relation to adjudications subject to a comprehensive self-contained statutory code, the presumption, in my opinion, must be that where the statute has created a specific jurisdiction for the determination of any issue which establishes the existence of a legal right, the principle of res judicata applies to give finality to that determination unless an intention to exclude that principle can properly be inferred as a matter of construction of the relevant statutory provisions."

202. I am unable to accept these submissions. To begin with the comments of Laskin J, they are obviously entitled to great respect, and it was of course Lord Radcliffe himself who described the Caffoor principle as "somewhat anomalous". But the principle is well established in English tax law; its continuing vitality was not doubted by the Court of Appeal, or by Lord Hope and Lord Hoffmann, in the Westmoreland case; it has been applied at tribunal level in relation to VAT, and by the High Court in relation to excise duty in Matalan; and the ECJ in Olimpiclub has held that res judicata and the principle of legal certainty must yield to the EU principle of effectiveness in the field of VAT, at least when the EU doctrine of abuse of rights is potentially in issue. Furthermore, if justification is needed for the truly anomalous features of the Caffoor principle (see paragraph 174 above), it may in my judgment be found in a combination of: (a) the public interest in taxpayers paying the correct amount of tax; (b) the EU principle of fiscal neutrality, in the sense that similar transactions should be taxed in the same way; and (c) the wider considerations of fairness between taxpayers which were identified by the US Supreme Court in Stone & Downer.
203. In Thrasyvoulou, the point actually decided by the House of Lords was that the principle of res judicata applies to decisions made by inspectors under the Town and Country Planning Act 1971. The presumption formulated by Lord Bridge at 289D is not an irrebuttable one, because he accepted that an intention to exclude the principle may be inferred as a matter of construction of the relevant statutory provisions. The House heard no argument on the Caffoor line of authority, and it was expressly accepted by counsel for Mr Thrasyvoulou that res judicata is not applicable in taxing and rating matters: see the report of their argument at 281E. The Caffoor principle is ultimately a principle of construction of the relevant tax legislation, and it is based on inferences drawn from the statutory scheme as a whole rather than any express provision. The question in the present case is whether a similar approach should be followed in relation to VAT. If (as I have held) an examination of the statutory context leads to the conclusion that the same principle should apply in relation to at least some aspects of VAT, including those in issue in the present case, I do not consider that such a conclusion is in any way precluded by Lord Bridge's presumption. Rather, it is an example of a case where the presumption is rebutted.
204. Mr Rabinowitz also referred me to two decisions at tribunal level which, he said, supported Littlewoods' argument. Carter Lauren Construction Ltd v Revenue and Customs Commissioners [2007] STC (SCD) 482 was a decision of a single special commissioner, Mr Charles Hellier, relating to the issue of construction industry scheme certificates under section 561 of the Income and Corporation Taxes Act 1988. In the context of appeals from the refusal of certificates brought under section 561(9), the special commissioner considered and accepted an argument advanced by the Revenue that an issue estoppel arose in relation to certain matters previously determined by the general commissioners. Although Mr Hellier's decision contains a thoughtful discussion of the question, the statutory context is so far removed from that of the present case, and other cases where the Caffoor principle has been held to apply, that I am unable to derive much assistance from it.
205. The second case, University College London [2008] UKVAT V20664, was a decision of the Tribunal released on 1 May 2008. Unlike Carter Lauren, it was a VAT case, the basic issue being whether the settlement of an earlier appeal by means of a section 85 agreement gave rise to an issue estoppel that precluded the Revenue from re-arguing the same point of principle that had arisen for determination in the first appeal. The Tribunal, chaired by Mr John Clark, accepted UCL's argument that issue estoppel applied, finding support for its conclusion in the approach which Mr Hellier had adopted in Carter Lauren. Again, I have not found this decision to be of much assistance. The appeal was under section 83(e) of VATA 1994, on a matter of principle relating to the allowance of input tax which was not tied to any particular period. It is therefore distinguishable from the present case, where the relevant appeals were all brought under section 83(p) or (t) and related to particular periods. In addition, there are some aspects of the Tribunal's reasoning which may not be reconcilable with the view which I take of the Caffoor principle. To the extent that this may be so, I can only say that I respectfully disagree.
206. Finally, Littlewoods submitted that the Caffoor principle can only be prospective in its operation, and that it cannot prevent an issue estoppel arising in respect of supplies made in periods prior to the relevant determination. In my judgment there is nothing in this point. I cannot discern any principled basis upon which the operation of Caffoor should be thus confined, and even Mr Rabinowitz was unable to suggest one.
207. I will now state my conclusions on this issue:

(1) In the light of my review of the law, I accept HMRC's second overriding contention in its broader form. I consider that the Caffoor principle applies to the underlying determinations of VAT and section 85 agreements in the present case, and that no issue estoppel can arise in relation to the separate claims for interest now advanced by the claimants so as to prevent HMRC from arguing that the VAT was in fact due as a defence to the claims. The position is in my judgment similar in all essential respects to that considered by Jacob J in King v Walden, which I respectfully think was correctly decided.

(2) If the above conclusion is wrong, I would accept HMRC's contention in its alternative, narrower, form, and hold that HMRC are not estopped from arguing that the VAT was in fact due save in relation to the specific quarterly periods and the specific companies covered by the earlier determinations and section 85 agreements.

(3) For the avoidance of doubt, my conclusion in either its broader or its narrower form still leaves open the question whether it would be an abuse of process to permit HMRC to argue that the VAT was due. Although Mr Swift appeared at times to question this proposition, he rightly accepted in his closing submissions that issue estoppel and abuse of process are analytically separate issues. His point was, rather, that if issue estoppel did in principle apply, but the case fell within the Arnold exception, then the question of abuse of process should also be decided in HMRC's favour for substantially the same reasons as brought the case within the Arnold exception.

(4) I have already held that the Arnold exception would not apply, if and to the extent that I am wrong in my view that issue estoppel is excluded by the Caffoor principle. Before coming on to abuse of process, however, I must first deal as briefly as I can with the technical arguments on issue estoppel which were addressed to me, on the footing that (contrary to what I have now held) the doctrine of issue estoppel applies in the usual way and is not excluded by the Caffoor principle.

(d) Issue estoppel: the technical issues

(i) The same issue

208. As I have explained (see paragraph 152 above), it is a requirement if an issue estoppel is to arise that the same question, or issue, should previously have been decided in proceedings between the same parties (or their privies). But what is meant by "the same issue", bearing in mind the distinction between cause of action estoppel and issue estoppel? The answer, as it was put by Lord Keith in Arnold at 105E (quoted in paragraph 153 above), is that issue estoppel may arise "where a particular issue forming a necessary ingredient in a cause of action" (my emphasis) has been previously litigated and decided. It is therefore not enough that the issue was only incidental or collateral to the earlier cause of action.
209. The point was made with typical clarity by Diplock LJ in Thoday v Thoday [1964] P 181 at 198:

"There are many causes of action which can only be established by proving that two or more different conditions are fulfilled. Such causes of action involve as many separate issues between the parties as there are conditions to be fulfilled by the plaintiff in order to establish his cause of action; and there may be cases where the fulfilment of an identical condition is a requirement common to two or more different causes of action. If in litigation upon such cause of action any of such separate issues as to whether a particular condition has been fulfilled is determined by a court of competent jurisdiction, either upon evidence or upon admission by a party to the litigation, neither party can, in subsequent litigation between one another upon any cause of action which depends upon the fulfilment of the identical condition, assert that the condition was fulfilled if the court has in the first litigation determined that it was not, or deny that it was fulfilled if the court in the first litigation determined that it was."

210. Translated into the context of VAT, the relevant causes of action in the earlier litigation may in my judgment be identified as:

(a) the VAT treatment under the Sixth Directive of the 2.5% element of TPP commission taken in goods by agents (i.e. the question in dispute between the parties in the proceedings which culminated in Littlewoods (CA));

(b) the VAT treatment under the Sixth Directive of the 10% element of TPP commission, whether taken in goods or in cash (that being the issue in the 10% Commission Appeal, which was settled by the 2004 section 85 Agreement); and

(c) the VAT treatment under the Second Directive of the 10% element of TPP commission taken in goods before 1978 (i.e. the subject matter of the 2008 section 85 Agreement).

211. The issues which HMRC now seek to re-open are identical to the issues which arose in (b) and (c) above, although they wish to rely on them as defences to a different cause of action (namely the claims for interest on overpaid tax). Furthermore, one of the issues decided by the Court of Appeal in Littlewoods (CA), as a fundamental part of its reasoning, was that the entirety of TPP commission taken in goods (and not just the 2.5% element) is properly to be treated as a reduction in the consideration paid by the agent for secondary goods, and is not consideration for services provided by the agent. This, too, is an issue which HMRC now wish to re-open in the light of Grattan (ECJ).
212. It seems to me, therefore, that all of the issues which HMRC now wish to reargue are the same as ones which were determined in the earlier proceedings, either as a necessary and fundamental part of the reasoning of the Court of Appeal in Littlewoods (CA) or by virtue of the two section 85 Agreements. The first of the three requirements of issue estoppel is therefore satisfied.
213. I also consider that this requirement should be regarded as satisfied in relation to all the periods which are now in issue, even though Littlewoods (CA) and the 10% Commission Appeal only concerned supplies made during particular periods (two periods in 1997 in the case of Littlewoods (CA), and the periods covered by the section 80 claim in the case of the 10% Commission Appeal; the point does not arise in relation to the GMAC Appeals, which covered all the periods before 1978 which are now in issue). I say this for two reasons: first, it is common ground that the agency commission arrangements in force across the Littlewoods and GUS groups were at all material times substantially identical; and, secondly, the evidence establishes that the arrangements considered by the Court of Appeal in Littlewoods (CA) were substantially the same as those now in issue.

(ii) Privity

214. An issue estoppel binds not only the parties to the earlier determination but also their privies in blood, title or interest. Privity of interest is the only relevant category in the present case. The modern approach to privity of interest stems from the discussion of the subject by Sir Robert Megarry V.-C. in Gleeson v J Wippell & Co Ltd [1977] 1 WLR 510 at 515C to 516D. After making the fairly obvious point that the word "interest" could not in this context have the meaning of "mere curiosity or concern", Sir Robert Megarry identified the core content of the principle at 515E-H:

"Second, it seems to me that the substratum of the doctrine is that a man ought not to be allowed to litigate a second time what has already been decided between himself and the other party to the litigation. This is in the interest both of the successful party and of the public. But I cannot see that this provides any basis for a successful defendant to say that the successful defence is a bar to the plaintiff suing some third party, or for that third party to say that the successful defence prevents the plaintiff from suing him, unless there is a sufficient degree of identity between the successful defendant and the third party. I do not say that one must be the alter ego of the other: but it does seem to me that, having due regard to the subject matter of the dispute, there must be a sufficient degree of identification between the two to make it just to hold that the decision to which one was party should be binding in proceedings to which the other is party. It is in that sense that I would regard the phrase "privity of interest". Thus in relation to trust property I think there will normally be a sufficient privity between the trustees and their beneficiaries to make a decision that is binding on the trustees also binding on the beneficiaries, and vice versa."

This passage was expressly approved as setting out the correct approach by Lord Bingham in Johnson v Gore Wood & Co [2002] 2 AC 1 at 32D-E.

215. Sir Robert went on to observe that, since privity takes effect whether the party to the earlier proceedings wins or loses, the matter may often be tested by considering the position if the result in the earlier proceedings had gone the other way. He commented at 516B that:

"Any contention which leads to the conclusion that a person is liable to be condemned unheard is plainly open to the gravest of suspicions. A defendant ought to be able to put his own defence in his own way, and to call his own evidence. He ought not to be concluded by the failure of the defence and evidence adduced by another defendant in other proceedings unless his standing in those other proceedings justifies the conclusion that a decision against the defendant in them ought fairly and truly to be said to be in substance a decision against him."

216. The principle has been considered very recently by the Court of Appeal in Resolution Chemicals Ltd v H Lundbeck A/S [2013] EWCA Civ 924, [2014] RPC 5, where the leading judgment (with which Longmore and Moore Bick LJJ agreed) was delivered by Floyd LJ. He pointed out, in [29], that Sir Robert Megarry's test embraces two concepts: first, the interest which the subsequent litigant has in the subject matter of the first action, and, secondly, the identity of the parties. After discussing these concepts, Floyd LJ summarised the matter as follows at [32]:

"32. Drawing this together, in my judgment a court which has the task of assessing whether there is privity of interest between a new party and the party to previous proceedings needs to examine (a) the extent to which the new party had an interest in the subject matter of the previous action; (b) the extent to which the new party can be said to be, in reality, the party to the original proceedings by reason of his relationship with that party, and (c) against this background to ask whether it is just that the new party should be bound by the outcome of the previous litigation."

217. The question whether companies in the same corporate group are to be regarded as privies of each other has been considered in a number of recent cases. In Special Effects Ltd v L'Oréal SA [2007] EWCA Civ 1, [2007] Bus LR 759, the Court of Appeal disapproved the view expressed by Sir Andrew Morritt C at first instance ([2006] EWHC 481 (Ch), [2006] RPC 849 at [55]) that "prima facie, each company in a group is to be regarded as the privy of every other company in the group unless it demonstrates the contrary". Delivering the judgment of the court, Lloyd LJ said at [82] that "we could not agree with so general a principle", which went further than had been necessary for the Chancellor's decision. Lloyd LJ added, however, that the Chancellor's decision might have been justifiable on a more limited and specific basis which formed part of his reasoning. In bare outline, the party to the original dispute was the group company which held registered marks and conducted proceedings relating to them, while the group company said to be its privy used those marks under licence from the first company. As Lloyd LJ said in [82]:

"If a corporate group such as L'Oréal chooses to arrange its affairs, no doubt for good reason, in such a way that matters such as trade mark oppositions, as well as applications and the holding of registered trade marks, are conducted by one company, for the benefit of others in the group, and others then use marks of which the first is the registered holder, or other marks, not yet registered, of which the first would be the holder if a registration was obtained, then it seems to us that it might well be consistent with what Sir Robert Megarry V-C said in [the Gleeson case] to regard any constraint on the first, whether by way of cause of action estoppel, issue estoppel or abuse of process, as applying also to the second as its privy."

218. In Resolution Chemicals, Floyd LJ commented at [47] that in the L'Oréal case there was "a real sense" in which the opposition proceedings had been conducted by the first group company on behalf of the second, and for its benefit. The second company had been using the mark in question at the time of the opposition proceedings, and plainly intended to continue doing so. It therefore "had a concrete interest in the outcome".
219. According to the researches of counsel, there is only one decided case in which the doctrine of privity has been considered in a tax-related context, namely Secretary of State for Business, Innovation and Skills v Potiwal [2012] EWHC 3723 (Ch), [2013] Lloyd's LR 124. The basic facts are summarised in the headnote:

"Mr Potiwal was the sole director and 40% shareholder of a company connected with the fraudulent evasion of VAT in 2006. He was in sole charge of its commercial activities. In January 2009 in a challenge to the company's VAT assessment by HMRC, the VAT Tribunal concluded on appeal that Mr Potiwal knew that the company was participating in the fraudulent evasion of VAT … Mr Potiwal was in charge of the conduct of the company's appeal. He gave all relevant instructions to the company's solicitors and counsel. He was its only witness of fact and was cross-examined during the proceedings. He had a strong financial interest as well as a reputational interest in the appeal coincident with that of the company. Following the proceedings with HMRC, the Secretary of State … applied to disqualify Mr Potiwal as a director under section 6 of the Company Directors Disqualification Act 1986. In the subsequent proceedings Mr Potiwal denied that he had the requisite knowledge and required the Secretary of State to re-litigate the issue."

The matter came before Briggs J (as he then was) on an application by the Secretary of State to strike out that plea on the grounds of issue estoppel or abuse of process. In relation to issue estoppel, the question of privity had to be considered in relation to both parties: was Mr Potiwal a privy of his company, and was the Secretary of State a privy of HMRC? Briggs J held that Mr Potiwal and his company were privies, but that the overlap in interest between HMRC and the Secretary of State was not quite sufficient to make them privies. He nevertheless held that it would be an abuse of process to require the Secretary of State to re-litigate the issue of Mr Potiwal's knowledge of the VAT fraud.

220. In deciding (with some hesitation) that HMRC and the Secretary of State were not privies in relation to the issue about Mr Potiwal's knowledge, Briggs J was influenced by two considerations in particular: see [20] and [21] of his judgment. First, since the effect of identifying two parties as privies is automatic, and gives rise to an estoppel which prevents the relevant dispute being revisited, regardless of the circumstances of the first trial, and of the outcome, the law should "be slow to recognise privity of interest between different persons". Secondly, the modern tendency, exemplified by Johnson v Gore Wood, is to treat res judicata as an aspect of the law of abuse of process, and it would "go against the grain of the development of the law about abuse of process to identify for the first time a new class of privity of interest between two very different arms of government pursuing different aspects of the public interest". In relation to the second of these reasons, it should be noted that in Virgin Atlantic Lord Sumption has stated (at [25]) that res judicata and abuse of process are "juridically very different", the former being a rule of substantive law, while the latter is "a concept which informs the exercise of the court's procedural powers". He described them as "distinct although overlapping legal principles with the common underlying purpose of limiting abusive and duplicative litigation".
221. In the light of these principles, issues of privity would arise in the present case in relation to any issue estoppels derived from Littlewoods (CA) or the 10% Commission Appeal. No points on privity are taken by HMRC in relation to the GMAC Appeal, including the assigned claims advanced by Shop Direct Group.
222. In relation to Littlewoods (CA), the claimants accept that at the time of the Court of Appeal's decision in 2001 the Legacy GUS Claimants had not yet joined the Littlewoods group, so for that reason (if no other) they could not be regarded as privies of the sole appellant, Littlewoods Limited. Littlewoods argue, however, that the six active claimants among the Legacy Littlewoods Claimants, i.e. the second to seventh claimants in the second action ("the six companies"), should be so regarded. In support of this argument, they say that the six companies were all wholly-owned indirect subsidiaries of Littlewoods Limited; their VAT affairs were centrally managed by Mr Mitchell; and as members of the Littlewoods VAT group, of which Littlewoods Limited was the representative member, they had a concrete interest in the outcome of the appeal.
223. For their part, HMRC accept that there is privity between companies within the same VAT group, but they argue that the privity could not extend beyond the two VAT periods covered by the original decisions in 1997 that led to the proceedings which culminated in Littlewoods (CA). Since I have already rejected the argument that any issue estoppels should be confined to the actual periods which were in issue in the proceedings, it must I think follow from HMRC's concession that the six companies should be regarded as privies of Littlewoods Limited in all the periods when they were members of the Littlewoods VAT group. That was the case during all the periods which are in issue in the present proceedings.
224. In relation to the 10% Commission Appeal, the position is less straightforward. The only party to the appeal was again Littlewoods Limited, which alone entered into the 2004 section 85 Agreement with HMRC. The claimants argue, however, that all of the active claimants in the present proceedings should be treated as privies of Littlewoods Limited, that is to say not only the six companies, but also (of the Legacy GUS Claimants) Reality Group Limited, Kay & Company Limited, Shop Direct Limited and Shop Direct Group (as assignee). Of the active claimants, the following were members of the Littlewoods VAT Group as at 20 October 2004: the six companies, Reality Group Limited and Shop Direct Group.
225. The matters relied upon by the claimants in support of the argument are broadly the same as those noted above in relation to Littlewoods (CA). First, all of the active claimants were under common ultimate control. As I have said, the six companies were wholly owned indirect subsidiaries of Littlewoods Limited; while the relevant Legacy GUS Claimants occupied a position in the group structure similar to that of Littlewoods Limited, all of them being under the ultimate common control of a Jersey-incorporated company called LW Corporation Limited. Next, the central management of the claimants' tax affairs by Mr Mitchell included the Legacy GUS Claimants from the day they joined the Littlewoods group in May 2003. Mr Mitchell's evidence shows that he initiated and managed all of the repayment claims that are the subject of the present proceedings, including the 10% Commission Appeal and the further associated claims referred to in paragraph 49 above. In his most recent (fourth) statement, Mr Mitchell says that all of the repayment claims relevant to the present proceedings were made "for the ultimate economic benefit of the Littlewoods group of companies as a whole". Thus, it is argued, the active claimants have for present purposes all been managed as if they were one company with a single directing mind. In addition, there is some evidence to support the inference that, although the 10% Commission Appeal was not formally designated as a test or lead case by the Tribunal, it was agreed on both sides to allow it to proceed to a hearing while the other associated claims were held in abeyance to await its outcome. In practice, therefore, success by Littlewoods Limited in the 10% Commission Appeal was a precondition for the successful prosecution of all the other claims. Finally, it is suggested that the point may be tested by supposing that Littlewoods Limited had lost the 10% Commission Appeal. In those circumstances, say Littlewoods, HMRC would certainly have objected if Mr Mitchell had sought to re-litigate the issue by lodging an appeal on behalf of another member of the Littlewoods group, even if that member was not also a member of the Littlewoods VAT group.
226. In their written submissions, counsel for HMRC take a number of technical points in answer to Littlewoods' argument. They submit, for example, that there could be no privity in respect of the assigned claims advanced by Shop Direct Group, even though Shop Direct Group formed part of the Littlewoods VAT group when the 2004 section 85 Agreement was concluded, because the section 80 claims for the GUS VAT group were originally made in June 2003 by Experian Finance Plc as its representative member, and the subsequent assignment of those claims to Shop Direct Group cannot alter the fact that the only privity in relation to those claims was between the members of the GUS VAT group at the time when the claims were first made. I do not propose to examine these arguments in more detail, however, because I am unable to accept HMRC's basic argument that privity is in principle incapable of extending beyond the membership of a VAT group. It seems to me that there is no warrant for a proposition in such a rigid form, which would not sit at all easily with the more flexible and fact-sensitive test laid down in L'Oréal and Resolution Chemicals for ascertaining privity within a corporate group.
227. In the particular circumstances of the present case, I think it would be commercially unrealistic to draw a distinction between any of the active claimants in relation to the issues in the 10% Commission Appeal, or between any of the Littlewoods or GUS companies on whose behalf those claims are advanced, because the claims were all centrally managed and coordinated for the benefit of the Littlewoods group as a whole by Mr Mitchell. Every company had a concrete interest in the outcome, because it had paid VAT on the allegedly unlawful basis and now sought to claim it back. I also agree with the claimants that it is helpful to test the point by assuming that the outcome of the appeal had gone the other way. HMRC would surely have said, in that event, that the Littlewoods group (including the companies acquired from GUS) had had their opportunity to litigate the issue, and they should all be treated as bound by the outcome, whichever (if any) particular VAT group they happened to belong to. Had I not accepted HMRC's overriding argument based on the Caffoor principle, I consider that such a stance by HMRC would, on the facts, have been amply justified.
228. For these reasons, I would answer the technical questions on issue estoppel in the claimants' favour and hold that, if issue estoppels can arise at all, they extend to all of the payments in issue in the present proceedings. This conclusion is, of course, subject to my prior conclusion that the doctrine of issue estoppel is in fact excluded in the present case by the Caffoor principle. In relation to the GMAC point, my conclusion is also subject to an argument on contractual estoppel to which I will now turn.

(3) Contractual estoppel: the 2008 section 85 Agreement

229. In his opening submissions, Mr Rabinowitz indicated that the claimants wished to run a new argument, hitherto unpleaded, to the effect that HMRC were contractually estopped by the terms of the 2008 section 85 Agreement from denying that the VAT in question was not due. Draft amendments to the claimants' re-re-amended reply were subsequently produced, which pleaded that by clause 1 of the 2008 section 85 Agreement HMRC had agreed with the GMAC appellants that they "are entitled to the Repayment Amounts and simple interest that has been paid" and, by necessary implication, that HMRC had agreed that the VAT in question had not been due. It was said that this agreement gave rise to a contractual estoppel, and that in support of the pleaded construction of the agreement reliance would be placed, if necessary, on HMRC's letters of 27 June and 1 September 2008. In the second of those letters, HMRC said they had taken the view, when repaying the relevant amounts to the GMAC appellants, that the relevant commission taken in goods "adjusts the value of the supply at the time the secondary supply is made rather than some time after the supply is made …".
230. Although Mr Swift at first reserved his position in relation to the proposed amendment, by the time he came to make his closing submissions he confirmed that HMRC did not object to it. I therefore gave permission for the amendments to be made, and the point became a further sub-issue for me to determine. When I asked Mr Rabinowitz whether there might be an element of overkill in running this additional argument, he explained that the purpose of doing so was to circumvent either or both of the Caffoor principle and the Arnold exception, if I were otherwise minded to hold that they applied. If I were satisfied that there was a contractual estoppel, submitted Mr Rabinowitz, that would be a conclusive bar to HMRC seeking to resile from the repayments of tax and simple interest which had been made to the GMAC appellants.
231. The 2008 section 85 Agreement was expressed to be made between "Littlewoods Ltd and others" as appellants and HMRC as respondents. It was headed "Partial Settlement Agreement", and provided (so far as material) as follows:

"WHEREAS

(A) By letters dated 10 September 2007 and 11 October 2007 and subsequently a letter dated 27 June 2008, the Respondents have rejected the Appellants' claims made pursuant to section 80 of the Value Added Tax Act 1994 for recovery of output tax said to have been overpaid (the "Decisions"), which claims were asserted by the Appellants in the letters dated 25 June 2002 and 23 June 2003 (and revised in letters dated 19 October 2006 and 31 October 2006) ("the Claims").

(B) The Appellants have appealed against those rejections by notices of appeal dated 4 October 2007 and 29 October 2007 ("the Appeals") …

(C) One of the issues raised by the Appeals and a subject of the Decisions is whether the Appellants are entitled to recover overpaid output tax in relation to the period 1 April 1973 to 31 December 1977 (inclusive) in respect of certain commissions paid by the Appellants in goods (or by way of a discount from the price of goods) to certain agents ("the GMAC Goods For Commission Issue").

(D) The Respondents have since reconsidered their position on the GMAC Goods For Commission Issue and now accept that the Appellants are entitled to repayment of the output tax that relates to those commissions.

(E) The parties have agreed that the principal amount to be repaid on this basis is £4,528,957 ("the Repayment Amount"), which amount has now been repaid with simple interest in the amount of £12,773,444.96 (a total repayment of £17,302,401.96) …

IT IS HEREBY AGREED between the parties that:

1. The Appellants are entitled to the Repayment Amount and simple interest that has been paid as set out above, and the Decisions shall be treated as varied to that extent.

2. The Appellants will advise the Manchester Tribunal Centre that the GMAC Goods For Commission Issue has been agreed between the parties and that the Appeals should be treated as allowed and the Claims accepted to the extent of that issue only but without prejudice to the other issues raised in the Appeals.

3. This Agreement is without prejudice to any right the Appellants may have (including any rights arising at law or in equity, under European law, or under any statutory provision) to claim further interest (on a compound basis and/or at a higher rate) in respect of the Repayment Amount.

4. …

5. This Agreement takes effect pursuant to section 85 of the Value Added Tax Act 1994."

The agreement was signed by the parties' solicitors, and dated 23 September 2008.

232. For reasons which I need not go into, Mr Rabinowitz accepted that the benefit of any contractual estoppel arising from the 2008 section 85 Agreement could be claimed only by those of the appellants which are Littlewoods Legacy Claimants, but not by GUS Legacy Claimants, and with the further exception of Littlewoods Limited itself (which is not an active claimant in relation to these early payments of VAT).
233. Although the agreement was expressed to take effect pursuant to section 85 of VATA 1994, and therefore had the same consequences as if the Tribunal had determined the relevant appeals in accordance with its terms, the new argument asserts that section 85 did not deprive the agreement of its ordinary status as a binding contract, and it was therefore capable of giving rise to a contractual estoppel. The relevant type of estoppel, submitted Mr Rabinowitz, is that discussed by the Court of Appeal in Springwell Navigation Corporation v J P Morgan Chase Bank and others [2010] EWCA Civ 1221, [2010] 2 CLC 705, at [143] and following. Put shortly, it is open to parties to agree that a state of affairs will be the basis of their contractual dealings with one another, even if they know that it is not the case. In such circumstances, an estoppel by convention precludes either party from denying the truth of the assumed state of affairs. As a general proposition, I have no difficulty with this principle (which gains added support from the recent decision of the Privy Council in Prime Sight Ltd v Lavarello [2013] UKPC 22, [2014] 2 WLR 84, which was not cited to me). I also have no difficulty in accepting that the relevant parties to the 2008 section 85 Agreement agreed, for good consideration, that the appellants were entitled to repayment of the relevant VAT and simple interest, in the sums which had been paid to them. Indeed, Mr Swift did not argue the contrary. But the question remains whether there is scope for a contractual estoppel to exist and operate in cases where the parties have chosen to enshrine their agreement in a section 85 agreement.
234. Mr Swift submits that a section 85 agreement has the consequences specified by section 85 itself, and not those that would otherwise arise if the matter simply rested in contract between the parties. Section 85 deems a decision of the Tribunal to exist, to which the Caffoor principle applies in the same way as it would to an actual decision of the Tribunal. There is no room for a contractual estoppel to co-exist with the deemed Tribunal decision, even if such an estoppel would have existed had the parties not used the machinery of a section 85 agreement.
235. In support of this submission, Mr Swift referred me to the decision of the Court of Appeal in Schuldenfrei v Hilton [1999] STC 821. The issue in that case was whether the Revenue and a taxpayer had "come to an agreement" in relation to an assessment to capital gains tax which was under appeal, within the meaning of section 54(1) of TMA 1970, the wording of which is materially similar to that of section 85(1). The taxpayer's case was singularly lacking in merit. After prolonged negotiations about his capital gains tax liability for the relevant year, the Revenue sent him an amended notice of assessment which apparently adjusted the original assessment by reducing it to nil. This was fairly obviously a mistake, but the taxpayer took no action in response to the notice and did not request repayment of the sums which he had paid towards satisfaction of his liability under the original assessment. Instead, when the error came to light some six months later, the taxpayer claimed that he and the Revenue had "come to an agreement" that the original assessment should be discharged, within the meaning of section 54(1). This contention was rejected by the special commissioners, by Neuberger J on the taxpayer's appeal to the High Court, and by the Court of Appeal.
236. Delivering the leading judgment in the Court of Appeal, Jonathan Parker J said at [42] that he found it difficult to envisage a situation in which an agreement which was effective under section 54(1) would not also be enforceable as a binding contract at common law; but he left open the question whether a section 54 agreement "must invariably meet all the requirements of the common law in relation to the enforceability of contracts, including in particular the requirement of consideration". He continued:

"The question which arises under s54(1) is whether the Revenue and the taxpayer have "come to an agreement" in relation to the assessment under appeal. If they have, then the subsection itself prescribes the consequences which are to follow from that agreement. Thus, the question whether a s54 agreement has been concluded has to be considered in a statutory, not in a common law, context."

He went on to say that, in addressing this question, common law concepts such as that of offer and acceptance were still of assistance in deciding whether the parties had "come to" an agreement.

237. Evans LJ emphasised the place of section 54(1) in the statutory scheme, at 832f-g:

"1. The statutory scheme is clear. The taxpayer is given notice of the inspector's assessment of the amount of tax due from him. Unless he appeals, the tax becomes payable when the time for appealing, 30 days, expires. This is a statutory debt and no question arises as to a contract or agreement to pay. If he does appeal, then the amount of tax is determined by the outcome of the appeal; but if the parties "come to an agreement" whilst the appeal is pending, then what they agree takes effect as the determination of the appeal (see s54(1)). No other outcome is provided for, and so the appeal remains pending and the Special Commissioners have jurisdiction unless and until an agreement is reached.

2. In this context, therefore, an agreement has statutory consequences. It does not make the taxpayer's liability to pay contractual, but the agreement defines the amount of the statutory debt."

238. Mr Swift relies on these passages as showing that, where the parties have come to an agreement within the meaning of the section, the consequences which flow from the agreement are those prescribed by the section. Common law principles may help in ascertaining whether the conditions for the section to apply have been satisfied, but once the section is engaged, the position is governed exclusively by the statute. Mr Swift could, I think, have found some further support for this submission in the fact that the section does in various respects modify the ordinary law of contract. Thus, apart from the fact (confirmed in Schuldenfrei) that the parties may "come to an agreement" for the purposes of the section in circumstances where one or more of the requirements for a contract at common law are not satisfied, subsection (2) enables the appellant taxpayer (but not the Revenue) to resile from the agreement where he gives notice in writing to that effect within 30 days from the date of the agreement, and subsection (3) provides that where an agreement is not in writing the section shall not apply unless the fact that an agreement was come to, and the terms agreed, are confirmed by notice in writing given by the Revenue to the appellant, or vice versa. These modifications of the ordinary law of contract tend to suggest, in my view, that the statutory scheme of section 85 was intended by Parliament to be exhaustive for agreements which fell within its scope.
239. In reply, Mr Rabinowitz argued that the true position is, rather, that where a section 85 agreement is also a contract at common law, the ordinary law of contract continues to apply to it except to the extent that it is modified by the section. He supported this submission by reference to the decision of Popplewell J in R v Inspector of Taxes, ex parte Bass Holdings Ltd [1993] STC 122, 65 TC 495, where it was held that a section 54 agreement was susceptible to rectification in order to correct a double deduction in respect of group relief in the computation which formed the basis of the agreement. Popplewell J said at 132h:

"I see no reason why the ordinary law of contract should not apply to this agreement as to any other agreement. The effect of a concluded agreement under s54(2) is that it shall be final and conclusive but that does not mean that the court is not entitled to look and see whether all the ingredients necessary to the formation of a proper contract have been complied with. Thus capacity, fraud, mistake and such like matters seem to me to be available to a party who seeks to challenge the agreement on one or more grounds. In the sense that the agreement is res judicata of the issues which it determines it is clearly final and conclusive. But that does not mean, in my judgment, that the ordinary rules governing the formation of the contract are deemed to have been complied with."

240. I see no reason to question the actual decision in the Bass Holdings case. It would be strange if the court had no jurisdiction to rectify a written section 54 or section 85 agreement; and the equitable doctrine of rectification is not, of course, confined to documents which either contain, or evidence, contracts at common law. The remedy extends to written documents generally, including unilateral ones. But this point does not provide me with much assistance in relation to the problem I now have to consider, where the focus is on the consequences of an agreement which admittedly falls within the section.
241. Although the point is not free from doubt, I think the better view is that the consequences laid down by the section were intended by Parliament to be exhaustive. The section says that "the like consequences shall ensue for all purposes as would have ensued if … a tribunal had determined the appeal in accordance with the terms of the agreement" (my emphasis). The words "for all purposes" could hardly be more general. If, in the present case, the relevant appeals had been determined by the Tribunal, the determination would not have been a matter of contract between the parties. The Caffoor principle would in my view have operated, and there would in any event have been no contract which could ground a contractual estoppel. I do not believe that Parliament contemplated a hybrid world in which a contractual estoppel could co-exist with the ordinary consequences of a deemed determination by the Tribunal.
242. For these reasons, I prefer the submissions of the Revenue on this sub-issue and accordingly conclude that the contractual estoppel argument fails.

(4) Abuse of process

243. I come finally to the question whether the Revenue should be prevented from re-litigating the underlying tax issue on the ground of abuse of process. It was common ground that this question falls to be answered with primary reference to the well-known principles stated by Lord Bingham, after a review of the authorities, in Johnson v Gore Wood & Co, loc. cit., at 31A-F:

"… Henderson v Henderson abuse of process, as now understood, although separate and distinct from cause of action estoppel and issue estoppel, has much in common with them. The underlying public interest is the same: that there should be finality in litigation and that a party should not be twice vexed in the same matter. This public interest is reinforced by the current emphasis on efficiency and economy in the conduct of litigation, in the interests of the parties and the public as a whole. The bringing of a claim or the raising of a defence in later proceedings may, without more, amount to abuse if the court is satisfied (the onus being on the party alleging abuse) that the claim or defence should have been raised in the earlier proceedings if it was to be raised at all. I would not accept that it is necessary, before abuse may be found, to identify any additional element such as a collateral attack on a previous decision or some dishonesty, but where those elements are present the later proceedings will be much more obviously abusive, and there will rarely be a finding of abuse unless the later proceeding involves what the court regards as unjust harassment of a party. It is, however, wrong to hold that because a matter could have been raised in earlier proceedings it should have been, so as to render the raising of it in later proceedings necessarily abusive. That is to adopt too dogmatic an approach to what should in my opinion be a broad, merits-based judgment which takes account of the public and private interests involved and also takes account of all the facts of the case, focussing attention on the crucial question whether, in all the circumstances, a party is misusing or abusing the process of the court by seeking to raise before it the issue which could have been raised before. As one cannot comprehensively list all possible forms of abuse, so one cannot formulate any hard and fast rule to determine whether, on given facts, abuse is to be found or not … While the result may often be the same, it is in my view preferable to ask whether in all the circumstances a party's conduct is an abuse than to ask whether the conduct is an abuse and then, if it is, to ask whether the abuse is excused or justified by special circumstances. Properly applied, and whatever the legitimacy of its descent, the rule has in my view a valuable part to play in protecting the interests of justice."

244. Lord Bingham went on (at 32H-33A) to reject a subsidiary argument that the rule in Henderson v Henderson did not apply to Mr Johnson since the first action against his company had culminated in a compromise and not a judgment:

"An important purpose of the rule is to protect a defendant against the harassment necessarily involved in repeated actions concerning the same subject matter. A second action is not the less harassing because the defendant has been driven or thought it prudent to settle the first; often, indeed, that outcome would make a second action the more harassing."

Thus it is no obstacle to the potential application of the rule in the present case that the 10% Commission Appeal and the GMAC Appeal were resolved by the 2004 and 2008 section 85 Agreements. See too the observations of Lord Millett, to similar effect, at 59C.

245. Precisely because the doctrine of abuse of process may apply where the earlier proceedings were settled, without being the subject of an adjudication, Lord Millett drew a distinction (at 59D-F) between the doctrine of res judicata in all its branches, which (he said) may properly be regarded as a rule of substantive law, applicable in all save exceptional circumstances, and abuse of process, which "can be no more than a procedural rule based on the need to protect the process of the court from abuse and the defendant from oppression". As I have already noted, this distinction was in substance adopted by Lord Sumption (with the agreement of all members of the Court) in Virgin Atlantic at [25].
246. I should also note a further passage in Lord Millett's speech, at 60C-D, where he said that "particular care" needs to be taken where the claimant in the second action is not the same as the claimant in the first, but his privy:

"Such situations are many and various, and it would be unwise to lay down any general rule. The principle is, no doubt, capable in theory of applying to a privy; but it is likely in practice to be easier for him to rebut the charge that his proceedings are oppressive or constitute an abuse of process than it would be for the original [claimant] to do so."

247. Although Lord Bingham expressed the view that the question whether a party's conduct is an abuse of process is a unitary one, to be answered by reference to all the circumstances, both sides approached the question in the present case on the footing that the answer to it was likely to turn on much the same considerations as were relevant in deciding whether the Arnold exception to issue estoppel would apply (assuming for this purpose that the ingredients of issue estoppel were otherwise satisfied, and that its application was not excluded by the Caffoor principle). In particular, Mr Swift put Grattan (ECJ) at the forefront of his submissions, arguing that just as it was a decision of such significance as to engage the Arnold exception, so too it negated any abuse of process.
248. In my view this was the only basis upon which the Revenue could hope to justify what would otherwise be as clear-cut a case of abuse of process as one could readily imagine. It is enough to recall that, in addition to accepting the decision in Littlewoods (CA) and subsequently freely entering into the 2004 and 2008 section 85 Agreements, after first putting the Littlewoods group to the time, trouble, expense and anxiety of protracted litigation on the underlying liability issues, the Revenue then expressly conceded in the First Trial before Vos J that all of the VAT upon which interest is now claimed by all of the active claimants had been overpaid. Furthermore, this concession underlay, and was repeated in, the order for reference to the ECJ made by Vos J on 4 November 2010, as well as the United Kingdom's written observations submitted to the Court on 6 April 2011.
249. The description of the factual and legal background at the very beginning of the UK's written observations set the scene as follows:

"2. This case concerns the remedy given by national law where an undertaking has overpaid VAT. VAT was levied by [HMRC] contrary to the requirements of EU Directives on VAT when relevant provisions of EU law (and national implementing legislation) concerning the taxable amount of supplies were not properly applied to certain supplies made by Littlewoods who are the claimants in the proceedings from which this reference is made …. The taxable amount of those supplies was overstated and VAT was overpaid. This pattern repeated itself throughout the period 1973 to 2004."

250. The trial before me is the second stage of the split trial originally ordered by Chief Master Winegarten on 28 April 2009. If it were not for the intervening decision in Grattan (ECJ), I am sure it would never have crossed the Revenue's mind to attempt to re-open the issue of liability which had been conceded at the First Trial, and upon which the reference to the ECJ had been premised. In the absence of very special circumstances, it would have been an obvious abuse of process of the type that no court would countenance. Thus everything depends on the impact of Grattan (ECJ). I will not repeat what I have already said about its relative insignificance to the new case that HMRC wished to advance when they applied on 14 March 2013 to re-amend their defences. In the light of my assessment of Grattan (ECJ), there can be only one answer to the question of abuse of process. All questions of liability were to be decided at the First Trial. That was the last occasion for any attempt by HMRC to revisit the question whether the VAT had been overpaid. The opportunity was not taken, and the point was conceded with full knowledge of its financial implications for the Treasury. The reference to the ECJ proceeded on the footing that the VAT had been overpaid. If finality in litigation is to mean anything, that was the end of the road on this particular issue.
251. As with the Arnold exception, I emphasise that my conclusion on this issue might have been different if Grattan (ECJ) were in truth a case which had changed the relevant legal landscape in such a way as to show that the underlying VAT had been due all along. In those circumstances, it is possible that the balance would have come down in favour of allowing HMRC to re-open the question, even at this very late stage. I do not think it profitable, however, to speculate further about what my decision would have been in such an event. It is enough to say that, as matters now stand, I am satisfied that it would be an abuse of the process of the court if HMRC were permitted to re-open the question whether the VAT was due. I would answer issue 2B accordingly.

(5) Summary of conclusions

252. In bare outline, my conclusions on this complex part of the case are as follows:

(1) There is no scope for issue estoppel to operate, because (a) the Caffoor principle applies, and (b) the present claims are all for interest, which is not an issue which has previously been determined.

(2) Alternatively, the Caffoor principle applies so as to prevent the operation of issue estoppel save in relation to the claims for interest on the particular amounts of VAT which were determined to have been overpaid, by specific companies for specific periods, in (a) Littlewoods (CA), (b) the 10% Commission Appeal, and (c) so much of the GMAC Appeal as was settled by the 2008 section 85 Agreement.

(3) If I am wrong about the application of the Caffoor principle, HMRC would be issue-estopped from defending all of the present claims on the ground that the VAT was due, and the Arnold exception to issue estoppel would not apply.

(4) Littlewood's separate contractual estoppel argument in relation to the 2008 section 85 Agreement must be rejected.

(5) Whether or not issue estoppel applies, it would in any event be an abuse of process if HMRC were permitted to defend the present claims on the ground that the VAT was due.

VII. The adequate indemnity issue

(1) Introduction

253. After these lengthy preliminaries, I can now at last come to the issue which lies at the heart of the case. What did the ECJ mean when it said, in paragraph 29 of its judgment, that the EU principle of effectiveness requires that the national rules for calculation of interest (i.e. in the present context simple interest calculated in accordance with section 78 of VATA 1994) "should not lead to depriving the taxpayer of an adequate indemnity for the loss occasioned through the undue payment of VAT?" Once the nature of the test has been elucidated, it is then necessary to ask whether it has been satisfied by the payment of simple interest at the statutory rate to the claimants.
254. The agreed formulation of issue 2C reads as follows:

"Unless issue 2A is answered such that the VAT payments were all due and issue 2B is answered such that the Commissioners are entitled to defend all of the claims on that basis, does/did section 78 of VATA 1994 satisfy the requirements of the principle of effectiveness and, in particular, does/did that section deprive the claimants of an adequate indemnity for the loss occasioned through the undue payment of VAT? "

255. As I have already explained (see paragraph 15 above), in Portfolio Dividends (No. 2) I recently had occasion to consider the question whether HMRC are liable to pay compound interest on the successful claims by two companies in the Prudential Assurance group for the recovery of unlawfully levied ACT and corporation tax. In [194] to [204] of that judgment I set out some of the background to the question in this jurisdiction, including the judgments in Chalke (High Court) and Chalke (CA) which preceded the reference to the ECJ in the present case. I also referred to the preliminary discussion of the issue by Vos J in Littlewoods (No. 1) at [66] to [71] (much of which I quoted in [203]). In [205] to [210] I then set out the critical paragraphs from the judgment in Littlewoods (ECJ), and drew some conclusions from them.
256. To avoid the need for readers of this judgment to cross-refer to what I said in Portfolio Dividends (No. 2), I will now reproduce [205] to [208] of that judgment:

"205. After referring to the San Giorgio principle, the Court in Littlewoods ECJ continued as follows:

"25. The Court has also held that, where a Member State has levied charges in breach of the rules of Community law, individuals are entitled to reimbursement not only of the tax unduly levied but also of the amounts paid to that state or retained by it which relate directly to that tax. That also includes losses constituted by the unavailability of sums of money as a result of a tax being levied prematurely (Metallgesellschaft, paragraphs 87 to 89, and Test Claimants in the FII Group Litigation, paragraph 205).

26. It follows from that case law that the principle of the obligation of Member States to repay with interest amounts of tax levied in breach of EU law follows from that law.

27. In the absence of EU legislation, it is for the internal legal order of each Member State to lay down the conditions in which such interest must be paid, particularly the rate of that interest and its method of calculation (simple or "compound" interest). Those conditions must comply with the principles of equivalence and effectiveness; that is to say that they must not be less favourable than those concerning similar claims based on provisions of national law (or arranged in such a way as to make the exercise of rights conferred by the EU legal order practically impossible (see, to that effect, San Giorgio, paragraph 12; Weber's Wine World, paragraph 103; and Case C-291/03 MyTravel [2005] ECR I-8477, paragraph 17).

28. Thus, according to consistent case law, the principle of effectiveness prohibits a Member State from rendering the exercise of rights conferred by the EU legal order impossible in practice or excessively difficult …

29. In this case, that principle requires that the national rules referring in particular to the calculation of interest which may be due should not lead to depriving the taxpayer of an adequate indemnity for the loss occasioned through the undue payment of VAT.

30. It is for the referring court to determine whether that is so in the case at issue in the main proceedings, having regard to all the circumstances of the case. In that regard it should be noted that it is apparent from the order for reference that, under the provisions of section 78 of the VATA 1994, the Commissioners paid Littlewoods interest on the VAT levied in breach of EU law. Pursuant to those provisions, Littlewoods received payment of simple interest, in accordance with the said provisions, in an amount of £268,159,135, corresponding to interest due over about 30 years, which amount exceeds by more than 23% that of the principal sum, which amounts to £204,774,763.

31. As for verifying whether the principle of equivalence has been complied with in the case at issue in the main proceedings, it should be noted that compliance with that principle requires that the national rule in question apply without distinction to actions based on infringement of EU law and those based on infringement of national law having a similar purpose and cause of action. However, the principle of equivalence cannot be interpreted as requiring a Member State to extend its most favourable rules to all actions brought in a certain area of law. In order to ensure compliance with that principle, it is for the national court, which alone has direct knowledge of the procedural rules governing restitution actions against the State, to determine whether the procedural rules intended to ensure that the rights derived by individuals from EU law are safeguarded under domestic law comply with that principle and to consider both the purpose and the essential characteristics of allegedly similar domestic actions. For that purpose, the national court must consider whether the actions concerned are similar as regards their purpose, cause of action and essential characteristics (see, to that effect, Case C-63/08 Pontin [2009] ECR I-10467, paragraph 45 and case-law cited).

…

34. In the light of the foregoing, the answer to the questions referred is that EU law must be interpreted as requiring that a taxable person who has overpaid VAT which was collected by the Member State contrary to the requirements of EU VAT legislation has a right to reimbursement of the tax collected in breach of EU law and to the payment of interest on the amount of the latter. It is for national law to determine, in compliance with the principles of effectiveness and equivalence, whether the principal sum must bear "simple interest", "compound interest" or another type of interest."

Paragraph 34 was then repeated in the formal ruling of the Court at the end of the judgment.

206. Some important points of principle emerge with clarity from this judgment. In the first place, the ECJ has repeated (in paragraph 25) the principle which it stated in FII (ECJ) I at paragraph 205 that the right to reimbursement of unlawfully levied tax extends to amounts paid to, or retained by, the State which "relate directly" to that tax. Such amounts are said to include (and, I infer, are therefore not confined to) losses representing the time value of prematurely levied tax. The generality of the principle thus exemplified is then reinforced by paragraph 26, which says (twice over, for good measure) that the obligation to repay unlawful tax "with interest" follows from the case law recited in the two previous paragraphs. I therefore take it as now settled that EU law requires interest to be paid when unlawful tax is reimbursed. In such cases, the payment of interest is not merely an ancillary matter for national law to determine. On the contrary, it is a substantive part of the right to a refund recognised in the San Giorgio line of cases, whether it is regarded as an amount retained by the State which relates directly to the unlawfully levied tax, or as a loss sustained by the person who paid the tax through the non-availability of the money. Either way, it is essentially a claim for the time value of the money, although whether that value is to be ascertained (in terms of domestic law) by measuring the State's unjust enrichment, or by quantifying the loss to the claimant, is not stated, and must therefore be a matter which is left to national law to determine.

207. Secondly, the judgment lends no support at all to the notion that EU law draws a distinction between cases where tax is levied prematurely and cases where tax is overpaid. The possibility of the Court drawing such a distinction was expressly raised by Vos J in Littlewoods (High Court) at [68], but the ECJ has instead affirmed paragraph 205 of FII (ECJ) I and recognised the general principle that unlawfully levied tax must be repaid with interest. In so ruling, the Court was in substance following the views expressed by Advocate General Trstenjak in paragraphs 29 to 30 of her Opinion, where she said that it would be illogical to distinguish between the two types of case, and that the rights of a taxable person to repayment of the unlawful tax and interest "are based on the provisions of EU law prohibiting the taxes levied".

208. Thirdly, however, the Court declined to rule that the right to interest means a right to compound interest. Paragraph 27 states explicitly that it is for the internal legal order of each Member State to lay down the conditions for payment of interest, including whether it should be simple or compound. The only requirements of EU law are that the national conditions must comply with the familiar principles of equivalence and effectiveness. As the Court explained, this means in general terms that the conditions "must not be less favourable than those concerning similar claims based on provisions of national law", nor must they operate "in such a way as to make the exercise of rights conferred by the EU legal order practically impossible". It is in the context of the requirement of effectiveness that the Court said, in paragraph 29, that the national rules on interest "should not lead to depriving the taxpayer of an adequate indemnity for the loss occasioned through the undue payment of VAT".

257. In the light of the argument which I have now heard in the present case, I adhere to the conclusions which I stated in [206] to [208], but I would take this opportunity to make one small correction. Referring to paragraph 26 of the judgment of the Court, I mistakenly accused it of having said "twice over" that the obligation to repay unlawful tax with interest follows from the case law recited in the two previous paragraphs. It is in fact clear that the words "from that law" at the end of paragraph 26 refer to the immediately preceding "EU law", and not to the earlier case law mentioned at the beginning of the paragraph. The paragraph reads inelegantly in the English version of the judgment, but the grammar is clear. Any possible doubt on the point is removed by reference to the original French text.
258. I wish to emphasise the crucial point that the right to interest on tax levied contrary to EU law has now been unambiguously recognised by the ECJ as a right conferred by EU law. Before Littlewoods (ECJ), there was real doubt on this question, depending on which strand of European case law was taken to be applicable. As Advocate General Trstenjak explained in her opinion (paragraphs 27 to 30), the Court in a series of mainly earlier judgments (exemplified by Case 26/74, Roquette Frères v Commission, [1976] ECR 677) had ruled that payment of interest was an ancillary matter to be settled by national law, including questions such as the date from which it must be calculated, the rate to be applied, and whether it should be simple or compound. In a second series of mainly more recent judgments, however, the Court had ruled that at least in some circumstances the taxpayer has a right to payment of interest on taxes levied in breach of EU law. This line of authority began with the Hoechst/Metallgesellschaft case, where the Court ruled that an award of interest was necessary under EU law where corporation tax had been levied prematurely in the form of unlawful ACT.
259. It is worth quoting what the Advocate General said about this second line of case law, because it lies behind and helps to explain the analysis adopted by the Court in paragraphs 24 to 26 of its judgment:

"29. This new line of case-law has been confirmed in Test Claimants in the FII Group Litigation and Test Claimants in the Thin Cap Group Litigation. Furthermore, both judgments give ample proof that the arguments developed in Metallgesellschaft and Others regarding advance payments of tax in breach of EU law can also be applied to cases where the levying of tax as a whole infringed EU law. This is also logical. As grounds for the interest claim under EU law as a result of advance payments of tax, the Court proceeds from the finding that because of the unavailability of sums of money as a result of a tax being levied prematurely, the taxpayer has suffered losses which are to be regarded as amounts retained by the Member State or paid to it in breach of EU law. Because the levying of taxes in infringement of EU law also gives rise to unavailability of sums paid until they are reimbursed, there is no evident reason to distinguish between the taxpayer's interest claim under EU law in the context of advance payments made in breach of EU law and such a right in the context of payments made in breach of EU law per se.

30. In the light of these considerations, Member States which have levied charges in breach of EU law must in principle, according to the Court's more recent case-law, both reimburse the charges levied in breach of EU law and pay interest in compensation for the unavailability of the sums paid. The taxable person therefore has a right to reimbursement of the charge and a right to payment of interest. Those rights enjoyed by the taxable person are based on the provisions of EU law prohibiting the taxes levied."

260. There can be no doubt, in my judgment, that the right under EU law to payment of interest on all kinds of unlawfully levied tax is now firmly entrenched in the jurisprudence of the ECJ. Not only was Littlewoods (ECJ) a decision of a Grand Chamber of 13 judges, but it has since been followed and applied by the Court in at least two cases, British Sugar and Irimie.

(2) British Sugar and Irimie

261. British Sugar (Case C-147/10) was one of three joined cases (the other two being Case C-113/10, Zuckerfabrik Jülich AG v Hauptzollamt Aachen, and Case C-234/10) in which the Fourth Chamber of the Court gave judgment on 27 September 2012, some two months after the judgment in Littlewoods (ECJ). The cases concerned the amounts of sugar levies paid between 2002 and 2006 pursuant to European Commission regulations which were subsequently held to be invalid. One of the questions raised in the British Sugar case, on a reference for a preliminary ruling by the Chancery Division of the High Court, related to the payment of interest on the overpaid production levies. It was argued by the Rural Payments Agency (an Executive Agency of DEFRA) that British Sugar should not be awarded interest on any repayment of overpaid levies, because any such repayment would be matched by a commensurate repayment by the Commission to the Rural Payments Agency pursuant to the EU's own resources system. On the assumption that there was no legal basis in the relevant EU legislation which would allow for the recovery of interest by Member States from the Commission in relation to such repayments, it was said that the same principle should apply to any repayment made by the Rural Payments Agency to British Sugar: see paragraph 31 of the judgment.
262. This argument was rejected by the Court. After referring to paragraphs 25 and 26 of its judgment in Littlewoods (ECJ), the Court continued:

"67. Thus, irrespective of the question whether the Member State may recover interest on the European Union's own resources – which is a question which does not have to be resolved in the context of the present case – individuals entitled to reimbursement of sums paid unduly in respect of production levies in the sugar sector determined on the basis of an invalid regulation are also entitled to payment of the interest on such sums.

68. Therefore, it is not permissible for a national court to use its discretion to refuse payment of interest on the sums charged by a Member State on the basis of an invalid regulation on the ground that that Member State could not reclaim the corresponding interest on the European Union's own resources."

These points were then repeated in the answer to the fourth question in paragraph 69, and in paragraph 3 of the formal ruling of the Court.

263. British Sugar thus affirms that interest is payable under EU law when unlawful levies are reimbursed, even if the national body which makes the repayment is itself unable to recover interest on the levies which it originally paid over to an EU institution, and even though (by virtue of their onward payment) the national body was not itself enriched by receipt of the levies. These features of the case emphasise the potency of the right to interest, and strongly suggest to me that it is now regarded by the ECJ as conceptually indistinguishable from the right to repayment of the principal amount of unlawfully levied tax itself.
264. The issue in Irimie (Case C-565/11, Irimie v Administratia Finantelor Publice Sibiu and Another, [2013] STC 1321), as reformulated by the Court in paragraph 16 of its judgment delivered on 18 April 2013, was:

"… whether European Union law must be interpreted as precluding a national system … which limits the interest granted on repayment of a tax levied in breach of European Union law to that accruing from the day following the date of the claim for repayment of that tax."

In 2007, Mrs Irimie had bought a car registered in Germany. In order to register it in Romania, she paid a pollution tax to the Romanian authorities in September 2008. On 31 August 2009, she brought an action in Romania seeking recovery of the sum she had paid as pollution tax together with statutory interest on that tax from the date of payment. In the light of subsequent ECJ case law, her right to recover the tax was in the end uncontroversial, but her claim to interest was prima facie curtailed by statutory provisions which had been interpreted as meaning that interest on sums repaid from public funds could run only from the day following the date of the claim for repayment (i.e. in her case 1 September 2009, about a year after the tax had been paid). The referring court doubted whether this provision of national law was compatible with EU law, and accordingly sought a preliminary ruling from the ECJ.

265. In holding that the curtailment of the right to interest did indeed infringe EU law, the Court repeated its analysis in Littlewoods (ECJ), in materially similar terms, concluding as follows:

"26. As regards the principle of effectiveness, that principle requires, in a situation of repayment of a tax levied by a Member State in breach of European Union law, that the national rules referring in particular to the calculation of interest which may be due should not lead to depriving the taxpayer of adequate compensation for the loss sustained through the undue payment of the tax (see Littlewoods Retail and Others, paragraph 29).

27. In this case, it must be found that a system such as that at issue in the main proceedings, which limits interest to that accruing from the day following the date of the claim for repayment of the tax unduly levied, does not meet that requirement.

28. That loss depends, inter alia, on the duration of the unavailability of the sum unduly levied in breach of European Union law and thus occurs, in principle, during the period between the date of the undue payment of the tax at issue and the date of repayment thereof."

266. It can be seen that the Court felt able to determine the issue for itself, and did not (as in Littlewoods (ECJ)) remit it to the national court. The reason for this, no doubt, is that the structure of the national legislation was such that it was systematically incapable of providing Mrs Irimie with "adequate compensation" for her loss, since interest would inevitably not be payable for the period between the date when the unlawful tax was paid and the date when she initiated proceedings for its recovery. As the Court said in paragraph 28, Mrs Irimie's loss depended, among other things, on the length of the period during which she had been deprived of her money, and that period began when the unlawful tax was paid.
267. It will be noted that in paragraph 26 of Irimie reference is made to "adequate compensation" for the loss sustained by the taxpayer, whereas the phrase used in paragraph 29 of Littlewoods (ECJ) is "an adequate indemnity". Nothing turns on this apparent distinction, however, because in each case the original French text of the judgment uses the same phrase, "une indemnisation adéquate".
268. It is also worth noting the opinion of Advocate General Wathelet, delivered on 13 December 2012. Although now an Advocate General, he had been a judge of the Court between 1995 and 2003, and was the juge rapporteur in Hoechst, a case which despite its seismic impact in the UK was decided by a Chamber of five judges. The Advocate General considered, and rejected, an argument advanced by the Kingdom of Spain that the principle of effectiveness would be infringed "only if the interest payment was so small that it would considerably limit, "render meaningless" or reduce to zero the right to payment of interest". To similar effect, the Portuguese Republic submitted that it was for Member States to lay down details of the amount to which the taxpayer was entitled by reason of the infringement of EU law, "provided that those details do not entail a substantial reduction in the amount to which the individual is entitled and cannot be regarded as an obstacle to the exercise of that same right" (paragraphs 16 and 17 of the opinion). In paragraph 22, the Advocate General recited the standard formulations of the EU law principles of effectiveness and equivalence, and continued:

"In addition, contrary to the arguments of the Kingdom of Spain and the Portuguese Republic set out in points 16 and 17 above, although the principle of procedural autonomy leaves it to the domestic legal system of each Member State to determine the procedural remedies for safeguarding rights which individuals derive from European Union law, that principle cannot have the consequence of restricting or undermining the substance of those rights."

269. He then discussed the two lines of authority which Advocate General Trstenjak had identified in Littlewoods (ECJ), commenting at paragraph 25 that the earlier approach "could be understood as leaving essentially to national law the question not only of the method of calculating the interest on the tax to be repaid, but also the question of the actual right of the taxpayer to receive interest". After referring to cases in the second line of authority, including Hoechst, FII (ECJ) I and Littlewoods (ECJ) itself, and to the requirement that the national rules should not deprive the taxpayer of an adequate indemnity, he said at paragraph 29:

"29. In my opinion, the right to interest representing an adequate indemnity for the loss occasioned through the undue payment of tax contrary to European Union law ranks equally, in consequence of the Littlewoods Retail and Others judgment, with the right to repayment of the tax and is therefore a subjective right derived from the legal order of the European Union. In my opinion, that subjective right necessarily entails the payment of interest from the date of payment of the tax. It is obvious that it is from that date, and not from any other subsequent date, that the taxpayer suffers a loss arising from the unavailability of the sums in question."

270. This passage makes explicit the proposition which seems to me to be implicit in both Littlewoods (ECJ) and the British Sugar case, namely that the right to interest representing an adequate indemnity "ranks equally" with the right to repayment of the unlawfully levied tax. The description of this right as "a subjective right" reads oddly in English, but must I think mean a personal or private right, representing the French "droit subjectif". I was referred in this connection by Mr Rabinowitz to F H S Bridge, The Council of Europe French-English Legal Dictionary, (1994), which under the entry "Subjectif" defines "droit subjectif" as "right conferred on or exercised by an individual; individual (personal, private) right (as opposed to droit objectif)".

(3) Adequate indemnity

271. At this stage, I need to fill in some more background which it is relevant to have in mind when interpreting the ECJ's judgment.
272. First, the Court was expressly asked by the High Court whether it would accord with EU law if a taxpayer who had overpaid VAT received only reimbursement of the principal sums overpaid, together with simple interest on those sums pursuant to national legislation such as section 78 of VATA 1994: see Question 1 in the order for reference, quoted in paragraph 21 of the judgment of the Court. Question 2 then asked, if the answer to the first question was negative, whether EU law requires payment of compound interest as the measure of the use value of the sums overpaid in the hands of the Member State and/or the loss of the use value of the money in the hands of the taxpayer. If the answer to both questions was negative, the third question then asked in general terms:

"… what must the remedy that EU law requires the Member State to provide include, in addition to reimbursement of the principal sums overpaid, in respect of the use value of the overpayment and/or interest?"

273. Secondly, the case advanced by Littlewoods in their detailed written observations submitted to the Court was not that EU law always required the payment of compound interest on overpaid tax, but rather that interest reflecting the use value of the money received should be paid in all cases where tax had been unlawfully collected contrary to EU law, in order to satisfy the principle of effectiveness, and that the Court should state this principle while leaving it to the national court to apply it in the varying factual circumstances of each case. So, for example, it was said in paragraph 94 that:

"… Littlewoods does not suggest that this Court needs to specify the rate of interest, or (if appropriate) the frequency of compounding, that will satisfy the principle of effectiveness. These are factual matters that will differ depending on the circumstances of particular cases. This Court can however rule on the principle that must be applied, viz that a remedy must be given in respect of use value that is commensurate with the benefit gained by the Member State. It is then for the national court to determine what that principle requires in particular cases.

95. Thus, the decision of the Court of Justice in Metallgesellschaft required that interest be paid, but it was not necessary for the Court to specify the rate or to rule on compounding. That was (properly) left for the national courts to determine."

274. Thirdly, the principle of effectiveness was a recurrent theme throughout Littlewoods' written observations, building on a submission that the San Giorgio right to repayment of unlawful tax was itself an expression of the principle. In footnote 36 to paragraph 47, this was said to be common ground because HMRC had contended at the First Trial that "San Giorgio is an expression of the principle of effectiveness" (the reference is to paragraph 36(a) of HMRC's skeleton argument, signed by the same team of counsel as have appeared before me). Reference was also made to San Giorgio (Case 199/82, [1983] ECR 3595) at paragraph 12 and to Case C-192/95, Société Comateb v Directeur Général des Douanes et Droits Indirects [1997] ECR I-165 at paragraph 20, together with the observations of Advocate General Tesauro in paragraph 11 of his opinion.
275. In my judgment there can be no doubt about the correctness of this proposition. It is enough to cite the following brief extracts from San Giorgio:

"12. In that connection it must be pointed out in the first place that entitlement to the repayment of charges levied by a Member State contrary to the rules of Community law is a consequence of, and an adjunct to, the rights conferred on individuals by the Community provisions prohibiting charges having an effect equivalent to customs duties or, as the case may be, the discriminatory application of internal taxes. Whilst it is true that repayment may be sought only within the framework of the conditions as to both substance and form, laid down by the various national laws applicable thereto, the fact nevertheless remains, as the Court has consistently held, that those conditions may … not be so framed as to render virtually impossible the exercise of rights conferred by Community law …

…

14. On the other hand, any requirement of proof which has the effect of making it virtually impossible or excessively difficult to secure the repayment of the charges levied contrary to Community law would be incompatible with Community law … "

276. Littlewoods buttressed their submissions by reference to a number of non-tax cases in the jurisprudence of the Court, including the well-known case of Marshall v Southampton and South West Hampshire Health Authority (No. 2) [1994] QB 126 (Case C-271/91, [1993] ECR I-4367) ("Marshall"). Miss Marshall had been dismissed by the respondent health authority on reaching the age of 62, the retirement age stipulated for women, although the retirement age for men in similar employment was 65. She sued successfully for sex discrimination, and the case was remitted to the industrial tribunal for compensation to be assessed. Before the hearing, the health authority paid her the maximum sum of £6,250 then permitted as compensation under section 65(2) of the Sex Discrimination Act 1975 as amended. The tribunal, however, considering itself bound by Community law to award adequate compensation, made an award of £19,405, which included £7,710 in respect of interest. In due course, questions were referred to the ECJ by the House of Lords asking, among other things, whether it was essential to the due implementation of article 6 of Council Directive 76/207/EEC (the Equal Treatment Directive) that her compensation should not be less than the loss which she had sustained by reason of the discrimination, and that it should include an award of interest on the principal amount. The ECJ then answered these questions in Miss Marshall's favour.
277. The reasoning of the Court was founded on the principle of effectiveness, as the following brief extracts from the judgment make clear:

"22. Article 6 of the Directive puts Member States under a duty to take the necessary measures to enable all persons who consider themselves wronged by discrimination to pursue their claims by judicial process. Such obligation implies that the measures in question should be sufficiently effective to achieve the objective of the Directive and should be capable of being effectively relied upon by the persons concerned before national courts.

…

24. … the objective is to arrive at real equality of opportunity and cannot therefore be attained in the absence of measures appropriate to restore such equality when it has not been observed. As the Court stated in paragraph 23 of the judgment in Von Colson and Kamann, cited above, those measures must be such as to guarantee real and effective judicial protection and have a real deterrent effect on the employer.

25. Such requirements necessarily entail that the particular circumstances of each breach of the principle of equal treatment should be taken into account. In the event of discriminatory dismissal contrary to Article 5(1) of the Directive, a situation of equality could not be restored without either reinstating the victim of discrimination or, in the alternative, granting financial compensation for the loss and damage sustained.

26. Where financial compensation is the measure adopted in order to achieve the objective indicated above, it must be adequate, in that it must enable the loss and damage actually sustained as a result of the discriminatory dismissal to be made good in full in accordance with the applicable national rules."

278. The Court went on, at paragraph 31, to say this with regard to the award of interest:

"31. With regard to the second part of the second question relating to the award of interest, suffice it to say that full compensation for the loss and damage sustained as a result of discriminatory dismissal cannot leave out of account factors, such as the effluxion of time, which may in fact reduce its value. The award of interest, in accordance with the applicable national rules, must therefore be regarded as an essential component of compensation for the purposes of restoring real equality of treatment."

279. The decision of the ECJ in Marshall is of interest for several reasons:

(1) First, it shows that in the context of article 6 of the Equal Treatment Directive the ECJ equated financial compensation which is "adequate" with compensation which makes good "in full" the loss and damage actually sustained by the claimant. In this respect, the Court was departing from the views of its Advocate General (Van Gerven) who had expressed the view that compensation could be adequate even if it was for less than the full damage sustained (see paragraph 17 of his opinion).

(2) Secondly, although Marshall is not referred to in the judgment of the Court in the present case, the Court will have had in mind the claimants' references to it in their written observations.

(3) Thirdly, the references in paragraphs 26 and 30 to "adequate" compensation and "adequate reparation" may have suggested the use of the phrase "adequate indemnity" in paragraph 29 of the judgment in the present case. As I pointed out in Portfolio Dividends (No. 2) at [209], the concept of an "adequate indemnity" appears to have no precursor in the case law of the ECJ in the field of taxation. It is therefore of some relevance to know that the ECJ has used similar language, with the connotation that "adequate" compensation must be "full" compensation, in a broadly comparable context where the provision of effective compensation for breach of a fundamental principle of EU law was in issue.

(4) Finally, it should be noted that there was no issue about compound interest in Marshall. What Miss Marshall wished to retain was the award of simple interest, under section 35A of the Supreme Court Act 1981 (as it then was), which had been made in her favour by the industrial tribunal. It would therefore be wrong to read the reference, in paragraph 31 of the judgment of the Court, to the award of interest "in accordance with the applicable national rules" being an essential component of compensation, as implicitly placing an upper limit on the amount of interest that might be recoverable in other circumstances.

280. Returning to the present case, the written observations submitted by the United Kingdom argued, in summary, that:

(a) all matters to do with interest are normally ancillary questions for the national court to resolve, subject to the principles of equivalence and effectiveness;

(b) cases of premature payment of tax contrary to EU law are distinguishable, because the interest is then "the very objective sought" by the claimant;

(c) the relevant rules of national law have not made it either impossible or excessively difficult for Littlewoods to exercise their right to recover the overpaid VAT;

(d) in a number of different contexts, payment of simple interest has been held to be compatible with EU law, or is provided for in EU secondary legislation; and

(e) an examination of the legislation in 13 other Members States indicates that in all but one (Sweden) the position is the same as in the United Kingdom, namely simple interest is payable both on recovery by taxpayers of taxes unduly paid, and on recovery by the tax authorities of taxes paid late.

281. In relation to the principle of effectiveness, the UK's position was encapsulated in paragraph 70:

"70. In the United Kingdom, simple interest is charged to taxpayers in respect of late payment of VAT. The use of simple interest is regarded as providing a means of compensating the tax collector for the passage of time. As already noted, simple interest is also used to compensate taxpayers seeking recovery of overpaid VAT for the passage of time. It cannot be said that the use of simple interest as opposed to compound interest makes it impossible or excessively difficult for taxpayers to recover overpaid VAT."

282. The arguments put forward by the European Commission were summarised by Advocate General Trstenjak in paragraph 11 of her opinion, to the effect that:

"… a remedy which provides only for reimbursement of the principal sums overpaid and simple interest on those sums is compatible with EU law in so far as that remedy provides adequate restitution or compensation for the loss of the use of the money and in so far as no more generous remedy is available, as a matter of national law, in relation to other taxes."

I observe that this paragraph provides another possible source for the concept of "adequate indemnity" adopted by the Court, and it does so in a context which may be thought to suggest that simple interest will often suffice for that purpose.

283. Finally, there can be no doubt that the Court was well aware of the potential quantum of Littlewoods' claims. The order for reference itself recorded that Littlewoods were claiming further sums amounting to some £1 billion in aggregate, calculated by reference to the compounded rates of interest applicable to UK Government borrowing from time to time over the period in question. Furthermore, this was recorded by the Advocate General in paragraph 8 of her opinion.
284. Although the Advocate General concluded that Littlewoods had a right under EU law to payment of interest, she nevertheless considered that this right had been satisfied by the payment of simple interest, and that this involved no breach of the principle of effectiveness. The nub of her reasoning on this point is contained in paragraph 34 of her opinion:

"34. It is settled case-law that the principle of effectiveness prohibits the Member States from rendering virtually impossible or excessively difficult the exercise of rights conferred by EU law. In the context of determining the detailed rules governing an interest claim stemming from EU law, a breach of the principle of effectiveness would therefore arise only if the interest were so low that it largely deprived the interest claim stemming from EU law of substance."

It was in that context that she noted (in paragraph 37) that the claimants had been paid simple interest in excess of £268 million, accrued over a period of around 30 years, which according to her exceeded the principal sum by more than 25% (for the correct figure, see paragraph 290 below). She said that in her view this payment of interest "readily complies with the principle of effectiveness".

285. Mr Swift submitted that the Advocate General's reasoning and conclusions were endorsed by the Court in paragraphs 27 to 30 of its judgment, but in my view that cannot be correct. If the Court really considered that the payment of simple interest to the claimants clearly complied with the principle of effectiveness, and that there would have been a breach of the principle only if the interest paid were so low that it largely deprived the interest claim of substance, it would surely have said so. It would not have remitted the question to the national court for determination. Furthermore, the Court would not have formulated the "adequate indemnity" test as one which the principle of effectiveness required the national court to apply. There is no hint of any such test in the Advocate General's own reasoning, and on any view it appears to set the bar significantly higher than a test which merely asks whether the interest claim has been largely deprived of substance.
286. In my judgment, the key to the problem must lie in a proper understanding of the nature and content of the right to interest under EU law. Only when that has been ascertained is it possible to answer the question whether payment of simple interest under section 78, and nothing more, renders exercise of the right impossible in practice or excessively difficult. The principle of effectiveness sets minimum standards for the enforcement and protection in national law of rights derived from EU law. It is a secondary or adjectival principle of EU law, in the sense that it only comes into play once the substance of the relevant EU right has been defined. The principle cannot be used to whittle down the right, or to give it a variable content in different Member States. Rather, it is concerned with ensuring the effective vindication of the full content of the right throughout the Union, while making due allowances for the differences in national law and procedure in Member States.
287. It is for this reason that I have placed so much emphasis on the recognition by the ECJ of the right to interest on unduly levied tax as a right conferred by EU law, which ranks equally with the right to repayment of the unlawful tax itself. This is the crucial step forward which the Court took in the present case, and which it has subsequently confirmed in British Sugar and Irimie. The earlier theory, that all questions relating to interest are normally for national law alone to determine, has now been decisively rejected. It seems to me that the Court crossed the Rubicon, in relation to interest, when it said in paragraph 26 that the obligation of Member States to repay "with interest" amounts of tax levied in breach of EU law is a principle which itself follows from EU law.
288. Mr Swift sought to deflect the force of this point by arguing that, although the right to interest must now be recognised as one derived from EU law, it is a right of a different nature from the right to repayment of unlawfully levied tax. The latter right, he submits, is a fundamental one, typically expressed by the Court (as in San Giorgio, paragraph 12) as being a consequence of, and an adjunct to, the rights conferred on individuals by provisions of EU law which prohibit unlawful charges to tax. Such a right can only be vindicated by repayment in full of the unlawful tax. By contrast, the right to interest is of a looser, more variable character, as is shown by paragraph 27 of the judgment which still follows earlier case law in saying that, in the absence of EU legislation, it is for the internal legal order of each Member State to lay down the conditions in which such interest must be paid, including the applicable rate and whether the interest should be simple or compounded. True, those conditions must comply with the principles of equivalence and effectiveness; but if the content of the right is so loose-textured, it is still possible for the national law of a Member State to limit the payment of interest to simple interest only.
289. Consistently with this approach, Mr Swift submitted that it is also open to a Member State to maintain a system which confines claimants to simple interest, even if there may be occasional hard cases where simple interest amounts to much less than the loss of the use value of the money computed on normal principles. Mr Swift drew an analogy with statutes of limitation, which (provided they are reasonable) are well established to be compatible with EU law, even though their effect may be to prevent a person from making an otherwise valid claim under EU law (including a claim to repayment of unlawfully levied tax). Furthermore, limitation rules can operate in a harsh and arbitrary manner, depending on the precise date when a claim is made. If the claim is made one day before the expiry of the period, it can be pursued in full, whereas if it is made two days later, it will fail in its entirety.
290. At a simpler level, Mr Swift argued that if the ECJ had intended the right to interest to give the taxpayer a full indemnity for his loss in all cases, it would have used the adjective "full" rather than "adequate" in paragraph 29, and would certainly not have gone out of its way (as it did) to record in paragraph 30 the amount which Littlewoods had received as simple interest, and the proportion which that simple interest bore to the principal sum. (Incidentally, the Court's arithmetic in paragraph 30 is faulty: the interest exceeds the amount of the principal sum by about 31%, not 23%).
291. These are cumulatively powerful submissions, and to a limited extent (as I shall explain) I can accept them. But they do not, in my judgment, make sufficient allowance for the critical point that the right to interest is now derived from, and protected by, EU law, in the same way as the right to repayment of unlawful tax. The strength of this principle is demonstrated by the recent decisions of the ECJ in British Sugar and Irimie. In those circumstances, the concept of an adequate indemnity must in my view mean an indemnity (or compensation) which is at least broadly commensurate with the loss of the use value of the overpaid money in the hands of the taxpayer.
292. The relevant meanings of "indemnity" in the online version of the Oxford English Dictionary, 2^nd edition, are 3(a) ("Compensation for loss or damage incurred; indemnification") and 3(b) ("A sum paid by way of compensation"). The relevant meaning of "adequate" is 3(a) ("Fully satisfying what is required; quite sufficient, suitable, or acceptable in quality or quantity"). In certain contexts (such as school reports, or satisfaction questionnaires) "adequate" may be a term of faint praise, connoting the achievement of a middling, but barely satisfactory, standard. That is the Dictionary meaning 3(b): "Satisfactory, but worthy of no stronger praise or recommendation; barely reaching an acceptable standard; just good enough". In my judgment, however, that cannot be the meaning which the Court intended in paragraph 29, especially when coupled with the noun "indemnity".
293. If I am right thus far, it is clear that the approach and conclusion of Advocate General Trstenjak cannot be correct. Her test (see paragraph 284 above) sets the bar far too low, and is open to the fundamental objection that it seems to treat the principle of effectiveness as if it could itself somehow define or qualify the content of the EU right to interest which she had rightly recognised. The Court, by contrast, focuses on the loss occasioned to the taxpayer by the undue payment of VAT, and on the adequacy of interest to compensate him for such loss.
294. It is also clear, I think, that the loss which has to be indemnified is the loss of the use value of the money overpaid. In economic terms, that is the function of interest; and the Court was clearly aware that this was the type of loss which Littlewoods was seeking to recover. Indeed, the Court had recorded as a preliminary observation in paragraph 23 that Littlewoods were not bringing an action for compensatory damages based on infringement by the United Kingdom of EU law. Moreover, the Court's reasoning in paragraph 25 built upon earlier authority (in Hoechst and FII (ECJ) I) to the effect that a reimbursement claim extends to "losses constituted by the unavailability of sums of money as a result of a tax being levied prematurely". The Court will also have had in mind the European Commission's submission that the principle of effectiveness requires "adequate restitution or compensation for the loss of the use of the money".
295. The decision in Hoechst, relating to unlawful ACT, which the ECJ has always characterised as prematurely levied corporation tax, was implemented in the United Kingdom by a claim for use value, and in Sempra the House of Lords held that the claim could only be satisfied by an award of compound interest. Once it is recognised that EU law requires a similar remedy in all cases of overpaid tax (and not only when tax is paid prematurely), and that payment of compensatory interest is also a requirement of EU law, the conclusion may appear to follow inexorably: the only way to provide the taxpayer with adequate compensation for the lost use value of his money will be by an award of compound interest.
296. But if that is what the ECJ meant, why did it not simply say so? Why did the Court say in paragraph 27 that the choice between simple and compound interest is still a matter for the national court? And why, above all, did it expressly draw attention in paragraph 30 to the amount which Littlewoods had received as simple interest, and the proportion which that amount bore to the principal sum? These are puzzling questions, but I must do my best to answer them.
297. The reason why the ECJ did not say that, on the facts, Littlewoods were clearly entitled to compound interest is, I think, twofold. First, that is not how Littlewoods put their case as a matter of principle under EU law: see paragraph 273 above. Secondly, the Court may have wished to depart as little as possible from its standard learning that all questions to do with interest are for national courts to determine, subject to the principles of equivalence and effectiveness. The present case is not one, like Irimie, where there is a structural defect in the national legislation which means that adequate compensation for the lost use value of the taxpayer's money can never be given (except, perhaps, in rare cases where the claim for repayment follows almost immediately after the payment of the unlawful levy). By contrast, there will be many cases where simple interest under section 78 does arguably satisfy the "adequate indemnity" test. The difference between simple and compound interest is relatively insignificant for periods of up to around three or four years, particularly when interest rates are low. Most cases in the UK for the recovery of overpaid VAT will be for periods of four years or less, following the enactment of the limitation period now contained in VATA 1994 section 80(4) as amended. Thus the present case is of an exceptional nature, having regard to the length of time for which interest is payable, and the very high rates of interest which obtained during most of the first half of that period. As a general proposition, therefore, it is appropriate that the question whether section 78 provides an adequate indemnity should be left to the national court to determine, having regard to all the circumstances of the case.
298. It also needs to be remembered that the guidance given by the ECJ is not specific to the United Kingdom, but has to apply across the EU. Methods of calculating and awarding interest are likely to vary from Member State to Member State, as will the legal principles by reference to which loss of use value of money is ascertained. It is therefore unsurprising that the ECJ chose to give its guidance in very general terms. Indeed, it could hardly have done otherwise.
299. This still leaves the question why, in the second limb of paragraph 30, the Court chose to descend into particulars and to draw express attention to the amount already paid to Littlewoods by way of simple interest. I confess that I have no satisfactory answer to this question, and neither side was able to suggest one which to my mind carried any degree of conviction. I readily accept that, when remitting a question to the national court, the ECJ will sometimes make an observation about the facts which gives a strong indication of what it thinks the answer probably should be. But unless I have wholly misunderstood the preceding analysis and conclusions of the Court, it is just not credible to suppose that simple interest for a period of over 30 years could have been thought by the Court to indemnify Littlewoods adequately for the loss of the use value of their money. Perhaps the Court was doing no more than reminding the national court that this is indeed an extraordinary case in terms of quantum, and that credit must of course be given for the enormous sums that Littlewoods have already recovered as simple interest when quantifying their loss of use claim.
300. I would add, incidentally, that neither side suggested that a further reference be made to the ECJ, seeking clarification of the guidance which it has given. The Grand Chamber has spoken, and it is now the duty of national courts to interpret and apply that guidance as best they can.
301. The focus in paragraph 30 on "all the circumstances of the case", and the amounts already paid to Littlewoods, does at least help to show that one of HMRC's contentions cannot in my view be correct. The contention was that, at a generic level, section 78 is likely to provide adequate compensation for the great majority of taxpayers who have overpaid VAT. As a general proposition, that may hold good for the majority of cases brought after the introduction of the limitation period in section 80(3). The question is not before me, and I express no concluded view on it one way or the other. But paragraph 30 makes it clear, to my mind, that the question of adequate indemnity must be answered by reference to the particular facts of the case in which it arises. The most that could be said, in my view, is that the existence and terms of the statutory scheme for payment of interest are among the circumstances to which the national court must have regard. If it be the case that the statutory scheme does normally comply with the principle of effectiveness, that may be a good reason for saying that it should normally apply, and that the burden should be firmly on the taxpayer to displace it by showing that simple interest at the statutory rate will not provide him with an adequate indemnity for his loss of use value. It is also material to bear in mind, in this connection, that loss of use value is an inherently imprecise concept, and views may reasonably differ in a given case about how it should be quantified. In this respect a claim for loss of use value lacks the clear-cut certainty of a claim for repayment of the overpaid tax itself.
302. In sum, my overall conclusion on the difficult question of the meaning of the "adequate indemnity" test in paragraph 29 of the ECJ's judgment is that it requires payment of an amount of interest which is broadly commensurate with the loss suffered by the taxpayer of the use value of the tax which he has overpaid, running from the date of payment until the date of repayment.

(4) Has the payment of simple interest provided Littlewoods with an adequate indemnity for the loss of use value of the overpaid VAT?

303. I will deal with this question briefly, because HMRC did not advance any separate argument that it should be answered adversely to the claimants, at this stage of the analysis, if I interpreted the adequate indemnity test as requiring either a full indemnity for their loss or an indemnity commensurate with their loss. Nothing which I say here should be read as prejudging any of the issues on quantum which I will have to consider in due course. Rather, what I am now considering is the level of loss to Littlewoods against which the adequacy of the indemnity needs to be measured in order to found liability.
304. For this purpose, as indeed for the purposes of quantum generally, Littlewoods rely not, as one might logically expect, on the loss of use value of the overpaid VAT in their own hands, but rather on the use value of the money in the hands of the Government. Littlewoods accept that the former measure is conceptually the correct one, but they argue that the Government's loss of use value provides a convenient minimum measure of, or proxy for, their own loss. They rely on the unchallenged evidence of their expert, now (as before) the distinguished economist Professor John Kay CBE, that whatever the use value of the money may have been to the Government, it would have been significantly higher to the claimants.
305. Professor Kay's reasons for this conclusion, which I accept, are set out in paragraphs 48 to 52 of his report on quantum dated 5 July 2013. As he says:

"The rate at which the UK Government can borrow is, in virtually all conceivable circumstances, lower than the rate at which any commercial entity can borrow, and this has been true throughout the period of overpayments. This differential exists for two principal reasons:

(i) lenders perceive lending to government as carrying the lowest risk associated with any borrower; and

(ii) loans to the government are more liquid – i.e. easier to realise in a secondary market – than loans to other entities.

…

In my view, the cost of capital to the claimants would always have been higher – probably substantially higher – than the cost of government borrowing, and this was continuously true throughout the period from 1973 to 2012."

306. Professor Kay's evidence about the cost the Government would have incurred in borrowing the amounts of the overpayments over the periods of the claim is also substantially unchallenged. He explains his methodology in paragraphs 53 to 59 of his report. In summary, he takes account of the five main categories of Government debt (short, medium, long, money market and (since 1981) indexed), and the proportionate mix of those categories, or "pots", from year to year, and proceeds on the assumption that any incremental borrowing would have been made through a similar mix of similar instruments. None of this evidence was challenged in cross-examination, and HMRC's own expert, Dr James Richardson, who is Director, Fiscal and Deputy Chief Economic Adviser at HM Treasury, accepted that Professor Kay's approach was not unreasonable. Dr Richardson would himself have favoured using a simpler approach, using only ten-year gilts, but he accepted that this would produce somewhat higher figures than Professor Kay's. It seems to me probable that Professor Kay's approach is more accurate than Dr Richardson's alternative, and in the circumstances I accept Professor Kay's methodology and calculations without adjustment.
307. On that footing, and updating the figures to 31 October 2013 in accordance with Professor Kay's supplemental report dated 15 October 2013, the aggregate cost to the Government of borrowing amounts equal to the overpayments amounts to £1,480,805,504, which after deducting the simple interest paid leaves a net claim for compound interest of £1,212,646,369. Littlewoods accept, however, that for the purposes of the adequate indemnity issue it may be appropriate to take a different calculation period, because the relevant question is whether the payments of simple interest, at the times when those payments were made, provided an adequate indemnity for the claimants' loss. I agree with that submission. On that basis, Professor Kay has recalculated the figures down to the respective dates of the payments of simple interest. This revised calculation yields an aggregate cost of borrowing of £1,102,065,414, which after deduction of the simple interest paid leaves a net claim of £833,906,279.
308. There is a further issue whether, in assessing the claimants' loss, account should be taken of the additional corporation tax which they would, hypothetically, have paid if they had not made the overpayments of VAT. I will deal with this argument when I come on to quantum. It is enough to say, at this stage, that it is common ground that some additional corporation tax would have been paid, but Littlewoods argue that EU law prevents its being taken into account. If it may be taken into account, there is then a disagreement between the relevant experts (Mr David Prestwich on behalf of the claimants, and Mr Christopher Barker on behalf of HMRC) about how much extra corporation tax would have been paid. Mr Barker estimates that an additional £74 million (approximately) would have been paid, whereas Mr Prestwich's estimate is about £64 million. The difference between their estimates turns on the assumptions that it is appropriate to make about how the group would have managed its tax affairs with the benefit of the extra money, and the extent to which the companies in question would have paid an effective rate of corporation tax which was lower than the statutory rate. I can say at once that I preferred the evidence of Mr Prestwich on these questions, and accept his evidence of the additional tax that would have been paid.
309. Although the difference between the estimates of the two experts is only some £10 million, the difference becomes much more significant when the numbers are run through the interest calculations. Professor Kay has again done the necessary calculations, both to 31 October 2013 and to the dates of payment of simple interest. Taking the latter dates, and Mr Prestwich's figures for corporation tax payments avoided, the total cost of borrowing net of corporation tax is £808,270,624, which after deduction of the simple interest paid leaves a net loss of £540,111,489. If Mr Barker's figures for corporation tax are taken, the net claim becomes £429,576,669.
310. For present purposes, it does not matter which of these figures is correct, or whether account should be taken at all of the additional corporation tax. On the approach most favourable to HMRC, that is to say assuming a deduction for corporation tax payments avoided, and taking Mr Barker's figures, there is still a net loss of nearly £430 million as at the dates when simple interest was paid. It follows that the simple interest paid could not, on any view, have provided the claimants with an adequate indemnity for their loss, and I would hold accordingly.

VIII. Conforming construction or disapplication?

(1) Introduction

311. On the basis of my conclusions so far, I have answered the issue of liability under EU law in favour of the claimants, and held that their entitlement to interest under EU law is not satisfied by the payment of simple interest under section 78 of VATA 1994. The next question is how, in accordance with the EU principles of equivalence and effectiveness, the claimants' entitlement to interest is to be given effect in English law. In particular:

(a) can sections 78 and 80 of the 1994 Act be construed conformably with EU law, under the Marleasing principle?

(b) if not, to what extent do sections 78 and 80 need to be disapplied? And in either case,

(c) are the claimants confined (as Vos J provisionally held) to the pursuit of their Woolwich claims, or are they also free to pursue their mistake-based DMG claims?

312. These questions are wrapped up in agreed issue 3, which reads as follows:

"If issue 2 is answered in the affirmative, can sections 78 and 80 of VATA 1994 be construed so as to conform with EU law (and if so, how), or must they be disapplied so as to allow either (a) only the Woolwich claims, or (b) both the Woolwich claims and the mistake-based claims?"

313. Vos J dealt with these questions on a preliminary basis in [72] to [92] of Littlewoods (No. 1), while recognising that the matter would in any event need to be reconsidered after the ECJ had given its judgment on the reference. His provisional answers to the two questions in issue 3 (see [92]) were that:

(a) sections 78 and 80 cannot properly be construed so as to conform with EU law; and

(b) to give effect to the claimants' putative San Giorgio right to the use value of the overpayment of tax, the sections "must be disapplied so as to allow only the Woolwich claims".

314. Vos J began his discussion of this issue by observing that it is primarily a matter for English law, in the sense that EU law will have determined the extent of the San Giorgio right, and it will be a matter for English law to provide a suitable remedy to give effect to the right. He said, and I respectfully agree, that EU law will not be concerned with the precise nature of the national law remedy by which means this is achieved, so long as the principles of equivalence and effectiveness are satisfied. He referred to section 2 of the European Communities Act 1972, which requires English legislation "to be construed and have effect subject to" EU rights. He said, and again I agree, that it is clear, as a matter of English law, that English statutes must be construed so as to accord with EU law if at all possible, and only if that is impossible will they need to be disapplied to allow EU rights to be given effect. He referred to the succinct summary by Sir Andrew Morritt C in Vodafone 2 v Revenue and Customs Commissioners (No. 2) [2009] EWCA Civ 446, [2010] Ch 77, at [37] to [38] of the principles to be observed in looking for a conforming interpretation, and continued (at [74]):

"For present purposes, it is sufficient to highlight the constraints on the broad and far-reaching nature of the interpretative obligation, namely that the meaning should "go with the grain of the legislation" and be "compatible with the underlying thrust of the legislation being construed", and should not be "inconsistent with a fundamental or cardinal feature of the legislation" since this would cross the boundary between interpretation and legislation."

315. As I pointed out in Portfolio Dividends (No. 2) at [101], the formulation by Sir Andrew Morritt C of the principles of conforming interpretation has recently been restated by the Court of Appeal in a single long paragraph, which I reproduced: see Wilkinson v Fitzgerald [2012] EWCA Civ 1166, [2013] 1 WLR 1776, at [50]. I also referred, in [102], to the comments of Lord Sumption in FII (SC) where at [176] he described the Marleasing principle, as it has been applied in England, as "authority for a highly muscular approach to the construction of national legislation so as to bring it into conformity with the directly effective Treaty obligations of the United Kingdom".
316. The relevant provisions of VATA 1994, which (as construed by myself in Chalke (High Court) and by Vos J in Littlewoods (No. 1)) prevent the claimants from advancing their claims for loss of the use value of the overpaid tax at common law are the following:

"78(1) Where, due to an error on the part of the Commissioners, a person has –

(a) accounted to them for an amount by way of output tax which was not output tax due from him …

…

then, if and to the extent that they would not be liable to do so apart from this section, they shall pay interest to him on that amount for the applicable period, but subject to the following provisions of this section.

…

80. …

(7) Except as provided by this section, the Commissioners shall not be liable to credit or repay any amount accounted for or paid to them by way of VAT that was not VAT due to them."

(2) Conforming construction

317. Against this background, the contention which Mr Rabinowitz placed at the forefront of his submissions on this part of the case was a beguilingly simple one. He pointed out that section 78(1) does not contain an absolute prohibition on recovery of interest outside the scope of the section, unlike section 80(7) which says in terms that HMRC shall not be liable to repay VAT which was not due "except as provided by this section". By contrast, section 78(1) provides for HMRC to pay interest "if and to the extent that they would not be liable to do so apart from this section". The question is therefore how far those saving words extend. Construing them in a purely domestic context, both Vos J and I have held that they refer only to other statutory sources of interest, and do not permit restitutionary claims at common law for loss of the use value of the overpaid tax. Broadly speaking, the reason why we both reached that conclusion, although not for precisely the same reasons, was that to allow scope for such common law claims would subvert, and be inconsistent with, the statutory scheme of sections 80 and 78. Mr Rabinowitz does not, at this level, seek to challenge that conclusion in a purely domestic context, despite the apparent width of the saving words in section 78(1). But, he says, when it is a question of applying the "highly muscular" principles of conforming construction, in order to prevent a breach of EU law, there is no reason why the saving words should not be given their natural wide meaning, and they should accordingly be construed as permitting claims such as those now advanced by the claimants. The effect, if this argument is right, is that the claims simply fall outside the scope of section 78.
318. This argument was not presented to me in Chalke (High Court), but it was advanced, for the first time, in Littlewoods (No. 1), where Vos J dealt with it as follows:

"75. Mr Rabinowitz raised a new argument in this case, somewhat ingeniously it might be said, bearing in mind the numerous occasions upon which these issues have been rehearsed. He submitted that s78(1) could easily be interpreted in a way that conformed to the putative EU right to the use value of money, by simply reading the exclusionary words in s78(1) [i.e. what I have called the saving words] as carving out restitutionary claims for interest as well as statutory ones. This interpretation did no violence to the words used and, so to speak, "ticked all the boxes" of the principles I have mentioned.

76. Ingenious though this approach is, it seems to me that it cannot work. For much the same reasons as I have given under issue 1, and as Warren J gave in John Wilkins, and as Arden LJ gave in Monro, the construction for which Mr Rabinowitz contends cuts straight across the grain of the legislation and is contrary to its fundamental or cardinal features. The legislation provided that taxpayers like Littlewoods should have only simple interest when they were repaid VAT upon an error being made by the Commissioners, and Mr Rabinowitz's construction gives them something much more. That would not be construction but legislation. One may ask rhetorically, what on earth would have been the point of s78 if its exclusionary words were to be construed as allowing a quite different common law interest remedy in every case? I have little doubt that, if there is an EU right to the use value of the money, that right can only be given effect by the disapplication of s78 in cases where the right exists (which one may note will be substantially all cases to which it applies)."

319. Mr Swift argued that it was no longer open to Littlewoods to run this argument at first instance, because it had already been decisively rejected by Vos J in the passage from his judgment which I have quoted. My initial inclination was not to accept this submission, on the ground that [72] and [92] of the judgment indicate that the whole of Vos J's discussion of issue 3 was intended to be provisional, and liable to be reconsidered in the light of the ECJ's ruling. On reflection, however, I think Mr Swift's point is a good one, and that Vos J intended his ruling on Littlewoods' new argument in [75] and [76] to be final. I say this because, when Vos J came to settle the terms of the order for reference, he framed question 4 in language which presupposes a disapplication of provisions such as sections 78 and 80 of VATA 1994, and when setting out the views of the referring court on that question (in paragraph 64 of the order) he said that the High Court's provisional conclusion on it was set out in [77] to [95] of the judgment. Paragraph [77] comes immediately after the rejection of Littlewoods' new argument, and begins:

"The question then arises as to how s78 should be disapplied."

320. It seems to me, therefore, that it is now too late for Littlewoods to resurrect this argument at first instance. In case I am wrong, however, I will briefly state my views on it.
321. Given the terms in which question 4 was framed, it is unsurprising that the ECJ had nothing to say on the question of conforming interpretation. It merely said, in paragraph 33 of its judgment:

"As is apparent from consistent case-law, when faced with a rule of law that is incompatible with directly applicable EU law, the national court is required to disapply that national rule, it being understood that that obligation does not restrict the power of the competent national courts to apply, amongst the various procedures of the internal legal order, those which are appropriate to safeguard the individual rights conferred by EU law …"

322. Mr Rabinowitz argued, however, that support for the new argument could be found in the way in which the Supreme Court dealt in FII (SC) with the Revenue's argument based on section 33 of TMA 1970. This point needs a little explanation. It arose in relation to the relatively few claims brought by the FII test claimants for the repayment of tax levied by assessment, namely corporation tax under Case V of Schedule D. Such claims prima facie fell within the scope of the corporate equivalent of section 33. As Lord Walker explained in FII (SC) at [116], the conditions for application of the section at the relevant time were fourfold: (1) it applied only to excessive tax charged by an assessment as a result of an error or mistake in a return; (2) there was a six-year time limit; (3) there was to be no repayment if the erroneous or mistaken return was in accordance with practice generally prevailing at the time; and (4) the repayment was to be such as the Revenue considered reasonable and just. On the face of it, any application under the section by the test claimants to recover unlawfully levied corporation tax would have been bound to fail, because the tax was paid in accordance with the practice generally prevailing at the time of payment. Since such a defence would clearly be incompatible with the claimants' San Giorgio claims to repayment, I held at first instance that the national legislation must yield to the principle of effectiveness, with the consequence that the section had to be disapplied, even though in a purely domestic context the section 33 regime excludes other remedies: see FII (High Court) at [437] to [439] and [449]. No argument was addressed to me that the exclusive application of the section, even to San Giorgio claims, could be preserved by a process of conforming construction.
323. In the Court of Appeal, the Revenue argued for the first time that a conforming construction was possible, and the argument was accepted by the court. It concluded that the prevailing practice condition "is to be read as subject to the limitation that it applies only if and to the extent that the United Kingdom can consistently with its Treaty obligations impose such a restriction": see FII (CA) at [261]. The court also held, at [264], that such a conforming construction would go with the grain of the legislation. In this way, the relevant claims were brought within the exclusive scope of section 33 with its six-year limitation period. A further consequence was that those claims which were in time could not be brought in the High Court, because they fell within the jurisdiction of the general or special commissioners.
324. The Supreme Court was unanimous in allowing the test claimants' appeal from those conclusions. Having explained the background, Lord Walker dealt with the point at [119]:

"119. I have grave doubts as to whether that interpretation does not go against the grain of the legislation, since the "practice generally prevailing" condition is of long standing and has always been regarded as an important safeguard for the public revenue. I am inclined to think that Mr Aaronson was right … to call it a "cardinal feature" of the legislation. In my view the Marleasing principle can be applied in a simpler and more natural way by not construing section 33 as impliedly setting itself up as an exclusive provision (which it did not do expressly, unlike section 80 of the Value Added Tax Act 1994). The test claimants submit that the application of Marleasing cannot rework section 33 in a way that serves any relevant purpose. But to read it as non-exclusive does not go against its grain. It would merely exclude an implication which is itself no more than a process of statutory construction."

325. To similar effect, Lord Sumption said at [204]:

"… [Section 33] confers a right subject to highly restrictive conditions to invoke what is essentially a discretionary power of the commissioners to grant a refund of overpaid tax. No one suggests on this appeal that such a limited remedy could possibly be enough in itself to satisfy the virtually unqualified obligation of the United Kingdom to provide an effective means of recovering tax overcharged contrary to EU law. This does not of course matter if it is an additional remedy as opposed to an exclusive one. There is certainly nothing in the provision which expressly excludes the availability of other causes of action at common law. If that is its effect, it must be by implication. In the ordinary way, such an exclusion might be implied, on the ground that where Parliament confers a restricted right of recovery, that must impliedly displace a corresponding right at common law which would be unrestricted. However, it is axiomatic that the courts cannot imply an exclusion of unrestricted rights of action at common law where that would be inconsistent with an overriding rule of EU law that an unrestricted right must be available. Section 33 cannot therefore be an exclusive right to recover tax overcharged contrary to EU law."

326. It can be seen, therefore, that the Supreme Court rejected the conforming construction adopted by the Court of Appeal, and held instead that the section should be read as non-exclusive in its application to San Giorgio claims, thereby leaving the test claimants free to pursue their restitution claims in the High Court. In concluding that this was a permissible application of the Marleasing principle, Lord Walker was influenced by the fact that, unlike section 80 of VATA 1994, section 33 was not an expressly exclusive provision. By parity of reasoning, submits Mr Rabinowitz, section 78 of VATA 1994 should be construed in such a way that San Giorgio claims for the use value of overpaid tax either fall outside its scope altogether, or at least come within the saving words in subsection (1).
327. The submission is an attractive one, but there are two reasons why I am unable to accept it. In the first place, section 78(1) must be read together with section 80(7) and its absolute prohibition on the recovery of overpaid VAT outside that section. In my view it would clearly go against the grain of section 78(1) to construe it as leaving room for a claim to interest at common law on a principal sum of overpaid VAT. Such a claim cannot have been within the contemplation of Parliament when section 78 was first enacted in 1991, at a time when the predecessor of section 80 was already in force (for the legislative history, see Chalke (High Court) at [57] to [60]). Secondly, I agree with Mr Swift that, even if a conforming construction of section 78(1) were possible, it would still be necessary to answer the question what has to be done by way of conforming construction in order to make section 78(1) compliant with EU law. That in turn would require consideration of essentially the same questions as were debated before Vos J on the assumption that sections 78(1) and 80(7) had to be disapplied. It is therefore hard to see how the conforming construction route could ultimately lead to a different conclusion from the disapplication route.

(3) Disapplication

328. I now move on to the question of disapplication, on the assumption that effect cannot be given to the claimants' San Giorgio interest claims by a process of conforming construction. On this footing, it is common ground that sections 78 and 80 have to be disapplied, and that the EU principle of effectiveness must be fully complied with. But there is fundamental disagreement about how far the disapplication needs to go. According to Littlewoods, once the statutory bar on their common law claims has been removed, effectiveness requires that they should be free to pursue both their Woolwich and their mistake-based DMG claims, since both causes of action are in principle available to them under English law, and they should be able to exercise the freedom of choice between concurrent remedies which English law normally permits. According to HMRC, on the other hand, the door should be opened only to the minimum extent necessary to comply with the principle of effectiveness. That objective would be fully achieved, they submit, by permitting the claimants to pursue their Woolwich claims alone; and this would also involve the minimum interference with the will of Parliament as expressed in sections 78 and 80.
329. This important question was fully debated before Vos J in Littlewoods (No. 1). He dealt with it in [77] to [92] of his judgment, and expressed his conclusions as follows:

"90. Taking all these factors together, it seems to me that the cause of action that most naturally and comprehensively gives effect to Littlewoods' San Giorgio right to the use value of the repayments is the Woolwich claim, which is applicable, as we now know it to be, to all cases where the government has exacted tax, which was not lawfully due, whether or not a demand has been made. The mistake-based restitutionary claims would only vindicate the San Giorgio right to the use value of money in some circumstances – namely where an appropriate mistake had been made. Such a claim would, therefore, be of less general application than the Woolwich claims. The rules to be disapplied are the same in this case, so that factor is neutral, but one would generally want to lean in favour of the minimum possible disapplication of domestic law.

91. The Woolwich claims here will be subject to an (EU law compliant) six-year limitation period which will reduce Littlewoods' claims for compound interest to a relatively small amount. On the other hand, if the mistake-based restitutionary claims were to be allowed, it is common ground that the limitation period must be extended under s32(1)(c) so as to allow the full extent of Littlewoods' claims. It might be suggested, therefore, that the approach that I have adopted is driven by a desire to protect the Commissioners from the major part of the claims to compound interest in this case. I should make it clear that I have ignored that factor, which seems to me, as a matter of law, to be entirely irrelevant.

92. For the reasons I have given, therefore, and on the assumptions that I have adopted above, I would provisionally answer the two questions in issue 3 as follows:

(i) ss78 and 80 of VATA 1994 cannot properly be construed so as to conform with EU law; and

(ii) to give effect to Littlewoods' putative San Giorgio right to the use value of the overpayment of tax, ss 78 and 80 of VATA 1994 must be disapplied so as to allow only the Woolwich claims."

330. More recently, I have had occasion to consider essentially the same question, argued by the same leading counsel on each side as in the present case, in the Investment Trust Companies ("ITC") litigation. The basic issue in that litigation was whether a final consumer of goods or services, who has borne the full economic burden of unlawfully levied VAT charged to him by his supplier, can bring a direct claim against HMRC to recover the unlawful tax, to the extent that he is unable to recover it from the supplier. The claimants were closed-end investment trusts, which had been charged VAT on investment management services rendered to them by management companies dating back to at least 1990. The unlawfulness of the VAT was not finally established until after the decision of the ECJ in Case C-363/05, J P Morgan Fleming Claverhouse Investment Trust Plc v Revenue and Customs Commissioners, [2007] ECR I-5517, [2008] STC 1180. The claimants then brought mistake-based DMG (but not Woolwich) claims against HMRC for the unlawful tax which (for various reasons) they had been unable to recover from the managers.
331. In Investment Trust Companies (in liquidation) v Revenue and Customs Commissioners [2012] EWHC 458 (Ch), [2012] STC 1150, ("ITC (No. 1)") I held (in short) that the claimants did in principle have valid restitutionary claims against HMRC, but as a matter of English law the claims were excluded by section 80(7) of VATA 1994. As in the present case, however, I also held that the claimants had good claims under EU law, which were to be classified as San Giorgio claims. In that context, the final question which I had to consider was whether those claims were nevertheless barred, on the ground that the only cause of action available to the claimants was an extended version of the Woolwich cause of action ("extended", because the claimants were not themselves taxpayers in relation to the Revenue: their obligation to pay the unlawful VAT had been a purely contractual one, between them and the managers). Since the claimants had pleaded their case only in mistake-based restitution, failure at this last hurdle would have left them without a remedy.
332. I discussed this question in [149] to [171] of my judgment, and set out the criticisms which the claimants made of the provisional conclusions reached by Vos J in Littlewoods (No. 1): see in particular [166] to [169]. I decided to defer a final decision on the question, however, until after the ECJ had given its judgment on the Littlewoods reference and the Supreme Court had given its judgment on the appeals from FII (CA). As I said in [170], I thought there was a real likelihood that these decisions would provide guidance at the highest level both on the scope of the Woolwich principle and on the requirements of the EU principle of effectiveness.
333. After the Supreme Court and the ECJ had delivered their respective judgments, the case was restored for further argument before me on 22 January 2013, and on 26 March 2013 I delivered a second judgment (Investment Trust Companies (in liquidation) v Revenue and Customs Commissioners (No. 2) [2013] EWHC 665 (Ch), [2013] STC 1129) ("ITC (No.2)") in which I held:

(a) that the Woolwich cause of action was not available to the claimants, in their capacity as final consumers, so the choice of remedies issue did not arise; but

(b) on the assumption that both Woolwich and DMG remedies were in principle available to them, the normal principle of freedom of choice between concurrent remedies under English domestic law should prevail.

334. In so concluding, I respectfully disagreed with the provisional conclusions of Vos J in Littlewoods (No. 1), which had drawn heavily on the reasoning of the Court of Appeal in FII (CA), and predated the decision of the Supreme Court. In my view, the reasoning of the majority in FII (SC) strongly supported the view that the claimants' argument was correct. The key passages in the judgments of Lord Hope, Lord Walker and Lord Reed on which I relied are set out in [16] to [20] of my judgment. The core of my reasoning is to be found in [46] and [47]:

"46. On that basis [i.e. on the assumption that the Woolwich cause of action was in principle available to the claimants], the next main question is whether Mr Swift is right in his submissions about the appropriate starting point, and the distinction which he draws between the position in the present case and the position in the FII litigation. In my judgment he is plainly right to draw attention to the fact that s80 of VATA 1994 is a statutory provision of long standing, which as a matter of national law provides an exclusive remedy for the recovery of overpaid VAT, and rules out any common law cause of action which might otherwise co-exist with it. In that respect, the position is clearly different from that which faced the FII test claimants. But does the distinction remain important once the exclusionary effect of s80 has been disapplied by EU law? My answer to that question is no. In my judgment, once the exclusionary rule in s80(7) has been over-ridden, the position is the same as it would be if common law causes of action had always been permitted to co-exist with s80, and in those circumstances no warrant can be found, in either English or EU law, for confining a claimant to only one such common law remedy, or for trying to identify the remedy which objectively provides the best fit for the claim. That was, in essence, what the Court of Appeal held should be done in FII (CA), but that approach has now been shown to be wrong by the majority in the Supreme Court. In short, once the exclusionary rule has been removed by force of EU law, I see no answer to the simple point that the normal principle of freedom of choice under English domestic law should be allowed to prevail.

47. In my view substantial support for this approach may be found in the judgments of the majority in FII (SC), including in particular the passages from the judgments of Lord Hope, Lord Walker and Lord Reed to which I have referred earlier in this judgment."

335. I went on to say that I found further support for my conclusions in the claimants' arguments founded on DMG, where the House of Lords had recognised the choice of domestic remedies available to the FII test claimants, despite the fact that EU law required disapplication of the principle in President of India v La Pintada Compania Navigacion SA [1985] AC 104 (HL) before their claims could succeed under English law as it was then understood: see [49].
336. Although I did not cite it in ITC (No. 2), it is worth referring to what Lord Walker said about the "European context" at the end of his discussion of what he termed "the cause of action issue" in DMG at [142]:

"142. I reach this conclusion without any heavy reliance on principles of EU law, except perhaps in relation to the Pintada principle (which I regard as uncertain in relation to the law of unjust enrichment). But the European context confirms my conclusion. The domestic court must give DMG an equivalent and effective remedy, and that would not be achieved, in my opinion, if recovery were limited so as to exclude an alternative concurrent remedy which would be available in a dispute between private citizens."

This approach is in my opinion entirely in line with the views of the majority in FII (SC), and is inimical to HMRC's argument that Littlewoods should be confined to the Woolwich remedy.

337. I would also emphasise, as Lord Walker's observations make clear, that it is not only the EU principle of effectiveness which is in play, but also the principle of equivalence. In FII (SC), both Lord Hope and Lord Reed were clearly of the view that it follows from the principle of equivalence that both grounds of action should be available in English law to vindicate a San Giorgio claim: see [21], [212] and [224]. I respectfully agree, and can see no principled basis for taking a different view merely because the question arises as a result of the disapplication of section 80(7).
338. Since I have recently dealt with this question in detail in my two ITC judgments, and since the claimants are content to support both my reasoning and my conclusions on the issue in that case, I do not propose to traverse the same ground again in this judgment. I must, however, refer briefly to the decision of the ECJ on the third FII reference, which was delivered on 12 December 2013, about a month after the conclusion of the hearing before me. As envisaged at the hearing (see paragraph 14(v) above), the parties then filed short written submissions on the judgment.
339. The ECJ decided that the enactment of section 320 of the Finance Act 2004, retrospectively and without any transitional arrangements, infringed the EU principles of effectiveness, legal certainty and the protection of legitimate expectations. The Court stated its conclusion (repeated in paragraph 1 of its formal order) as follows:

"49. In view of the foregoing considerations, the answer to the question referred is that in a situation in which, under national law, taxpayers have a choice between two possible causes of action as regards the recovery of tax levied in breach of EU law, one of which benefits from a longer limitation period, the principles of effectiveness, legal certainty and the protection of legitimate expectations preclude national legislation curtailing that limitation period without notice and retroactively."

In reaching this conclusion, the Court implicitly rejected the submission of the United Kingdom Government (recorded in paragraph 26 of the judgment) that curtailment of the limitation period for the DMG cause of action (referred to in the judgment as "the Kleinwort Benson cause of action") "did not in any way affect the limitation period applicable to the Woolwich cause of action, which, in itself, satisfies the principle of effectiveness and which Aegis [the relevant test claimant] was at all times entitled to use to recover tax levied in breach of EU law". The Court was to my mind clearly of the view that the three EU principles protected both causes of action alike, and it was no answer to the unlawful curtailment of the DMG cause of action to say that the test claimant was still left with the Woolwich cause of action.

340. The same point had been made explicitly, in paragraphs 45 to 50 of his opinion, by Advocate General Wathelet. He also said, at paragraph 53, that:

"… the guarantees attaching to the principle of effectiveness apply to every legal remedy which national law makes available to claimants for the reimbursement of taxes levied in breach of EU law (my emphasis)."

The decision of the Court is entirely consistent with the reasoning of the Advocate General, and in my judgment the Court must be taken to have endorsed it. The decision is not of direct application to the present case, because the issue arose in a context where the test claimants were in principle free to pursue both causes of action, and the question was whether one of them could be validly curtailed. There was no equivalent of section 80(7), which (unless disapplied) prevents the claimants in the present case from pursuing either cause of action. But the situation seems to me closely analogous, because in each case a provision of UK primary legislation has to be disapplied in order to enable San Giorgio rights to be vindicated, and the Court has rejected the argument that EU law is not engaged so long as the claimant is left with a single EU law-compliant remedy.

341. For all these reasons, I would answer the questions raised by issue 3 as follows:

(a) sections 78 and 80 of VATA 1994 cannot be construed so as to conform with EU law; and

(b) they must be disapplied so as to allow the claimants to pursue both their Woolwich claims and their mistake-based DMG claims.

IX. Change of position and exhaustion of benefits

342. As formulated before Vos J at the First Trial, issue 4 asked whether HMRC were entitled to deploy a change of position and/or an exhaustion of benefits defence to (a) the Woolwich claims and/or (b) the mistake-based claims, and (if so) whether those defences were made out in fact and to what extent. Vos J broke down this composite question into six sub-issues, which he dealt with separately. Before doing so, he introduced them in a passage of his judgment which it may be helpful to repeat:

"97. The Commissioners have rather "lumped together" their two defences of (a) change of position and (b) exhaustion of benefits, even though, in my judgment, they are, if connected, legally and factually distinct. It is perhaps useful at the outset to state briefly what they are:

(i) The change of position defence asserts that it would be inequitable for the Commissioners to be forced to repay the use value of the overpayment of tax, because they changed their position by irretrievably spending those payments within each fiscal year.

(ii) The exhaustion of benefits defence asserts that the Government's annual budget procedures mean that tax receipts are used up in each fiscal year, and no benefit is derived by the Government from those receipts thereafter. For that reason, the claimants' measure of recovery should be limited to the benefit of the use of the principal sums for a single fiscal year. This argument seems to me to relate as much to the assessment of the benefit obtained by the Commissioners from the use of the overpayment as it does to a properly so-called change of position defence. I will nonetheless deal with it under this issue, as that is how the parties have argued it.

98. Again, if my preliminary views on issue 2 are endorsed by the ECJ, these issues will not arise at all. In addition, if the ECJ finds in favour of a San Giorgio right to the use value of the overpaid tax, the matter will come back to this court, and the applicability of these defences will need to be reconsidered in the light of the ECJ's decision.

99. Nonetheless, both parties have agreed that I need to deal with the English law aspects of these defences, so as to provide an appropriate foundation for the reference to the ECJ. I think I also need to deal with the EU law question of whether change of position defences can, in principle, be available in answer to San Giorgio claims, since this, as it seems to me, is part and parcel of the proper approach to the principle of effectiveness."

343. I will now set out the six sub-issues identified by Vos J, and his answers to them as summarised in [149] of his judgment.

(1) Issue 4A: "Is a change of position defence available in English law to (a) a Woolwich claim, and/or (b) a mistake-based restitutionary claim?"

Answer: "A change of position defence is available in English law to a mistake-based restitutionary claim, but not to a Woolwich-based claim".

(2) Issue 4B: "Is an exhaustion of benefits defence available in English law to (a) a Woolwich claim, and/ or (b) a mistake-based restitutionary claim?"

Answer: "In so far as an exhaustion of benefits defence is a species of change of position, it is, as a matter of English law, available to a mistake-based restitutionary claim, but not to a Woolwich-based claim".

(3) Issue 4C: "Have the Commissioners made their change of position defence good on the facts to (a) the Woolwich claims, and/or (b) the mistake-based restitutionary claims?"

Answer: "The Commissioners' change of position defence to the mistake-based restitutionary claims (if the claims were available in law) and to the Woolwich-based claims (if the defence was available in law) could not succeed on the facts, because the Commissioners have failed to prove that the government increased its spending on the basis of the expectation or the happening of the overpayments, even though they have shown that they either spent or used the overpayments to reduce governments borrowing".

(4) Issue 4D: "Have the Commissioners made their exhaustion of benefits defence good on the facts to (a) the Woolwich claims and/or (b) the mistake-based restitutionary claims?"

Answer: "Whilst the Commissioners have not made any change of position defence good on the facts, whether framed as an exhaustion of benefits defence or not, it is open to the Commissioners to argue on the assessment of quantum in answer to either or both of the Woolwich-based and the mistake-based claims (if such were available) that the government has not benefited from the money after the first fiscal year, so that no use value of the overpayments should be awarded to Littlewoods in respect of a subsequent period".

(5) Issue 4E: "Assuming a change of position defence is available and has been made good on the facts, can it be given effect as a matter of EU law in answer to San Giorgio claims for the use value of overpayments of tax?"

Answer: "Assuming a change of position defence is available in national law, and has been made good on the facts, my provisional view is that such a defence should be given effect as a matter of EU law in answer to San Giorgio claims for the use value of overpayments of tax".

(6) Issue 4F: "Assuming an exhaustion of benefits defence is available and has been made good on the facts, can it be given effect as a matter of EU law in answer to San Giorgio claims for the use value of overpayments of tax?"

Answer: "In so far as the exhaustion of benefits defence is properly to be regarded as a change of position defence, my answer to this issue is the same as it is above for the change of position defence itself".

344. I heard no separate argument on any of these sub-issues, on the basis that they had either been finally decided by Vos J, or (where he left a question open, or said it could be revisited at the quantum stage) that it would be more conveniently dealt with in the context of the argument on quantum. I agree that this is the preferable course to adopt. When dealing with quantum, I will therefore refer where necessary to the views of Vos J when he discussed the relevant sub-issues. For the record, he dealt with issue 4A at [101] to [109]; issue 4B at [110] to [111]; issue 4C at [112] to [125]; issue 4D at [126] to [131]; issue 4E at [132] to [140]; and issue 4F at [141].

X. What is the measure of the claimants' recovery?

345. Issue 5 was formulated as follows by Vos J:

"In principle, is the measure of recovery for the Woolwich claims and/or the mistake-based claims to be measured by reference to (a) a conventional rate of interest compounded, or (b) the actual benefit enjoyed by the Commissioners, or (c) a compound rate of interest reflecting the cost of national government borrowing, or (d) in some other way? And is the measure limited to the lower of the value of (a) the use of which the claimants have been deprived and (b) the value of the use which the Commissioners received?"

346. Vos J discussed this issue in [142] to [147] of his judgment. His answer to the question, as summarised in [149], was this:

"(a) The measure of Littlewoods' Woolwich claim and its mistake-based claim (if one were available) for the use value of the overpayments of VAT is a compound rate of interest reflecting the cost of national government borrowing, but it is open to the Commissioners to show at the quantum hearing that they have benefited to a lesser extent so that they should only be liable for the actual benefit enjoyed by them.

(b) This measure of Littlewoods' claims is not limited to the lower of the value of (a) the use of which they have been deprived and (b) the value of the use which the Commissioners received."

347. It is common ground that these determinations were final, as far as they went, and again I heard no separate argument on them.

XI. Quantum

(1) The right approach to quantification of Littlewoods' claims

348. As I have already said, Littlewoods do not seek to recover the loss of use value to themselves of the overpaid tax, but rather the lesser amount representing the use value to the Government of the overpaid tax: see paragraph 304 above. They rely on the reasoning of the majority of the House of Lords in Sempra, and the expert evidence of Professor Kay, to argue that their claims should be quantified by an award of compound interest, computed in accordance with Professor Kay's "five pot" approach over the period from the dates when the tax was paid until judgment. The frequency of compounding, on this approach, is every six months for Professor Kay's four categories of Government stock, and quarterly for the money market pot (in line with the most common period to maturity of Treasury bills): see Appendix 3 to Professor Kay's main report dated 5 July 2013.
349. As an economist who has specialised in public finance and business economics, Professor Kay expresses his agreement with Lord Hope's statement in Sempra, at [33], that:

"Simple interest is an artificial construct which has no relation to the way money is obtained or turned to account in the real world. It is an imperfect way of measuring the time value of what was received prematurely."

He also agrees with Lord Nicholls' statement, at [52], that:

"We live in a world where interest payments for the use of money are calculated on a compound basis. Money is not available commercially on simple interest terms."

350. Speaking of Sempra's two restitutionary causes of action for recovery of the use value of the prematurely paid ACT, Lord Nicholls was at pains to analyse the issue from first principles:

"101. … Sempra's claim is that under both causes of action restitution requires the Inland Revenue to pay Sempra the value of the benefit the Inland Revenue obtained by having use of the money Sempra paid as ACT.

102. In principle this claim is unanswerable. The benefits transferred by Sempra to the Inland Revenue comprised, in short, (1) the amounts of tax paid to the Inland Revenue and, consequentially, (2) the opportunity for the Inland Revenue, or the Government of which the Inland Revenue is a department, to use this money for the period of prematurity. The Inland Revenue was enriched by the latter head in addition to the former. The payment of ACT was the equivalent of a massive interest free loan. Restitution, if it is to be complete, must encompass both heads. Restitution by the revenue requires (1) repayment of the amounts of tax paid prematurely (this claim became spent once set off occurred) and (2) payment for having the use of the money for the period of prematurity.

103. In the ordinary course the value of having the use of money, sometimes called the "use value" or "time value" of money, is best measured in this restitutionary context by the reasonable cost the defendant would have incurred in borrowing the amount in question for the relevant period. That is the market value of the benefit the defendant acquired by having the use of the money. This means the relevant measure in the present case is the cost the United Kingdom Government would have incurred in borrowing the ACT for the period of prematurity. Like all borrowings in the money market, interest charges calculated in this way would inevitably be calculated on a compound basis."

351. Lord Nicholls then explained how the then state of English law did not accord with this analysis, but went on to hold that the time had come to recognise that the court had power, in exercise of its common law restitutionary jurisdiction, to make an award of compound interest in order to achieve full restitution and bring about a just result: see [104] to [112]. He also explained, at [114], that awards of compound interest as restitutionary relief in respect of a defendant's unjust enrichment did not conflict with section 35A of the Supreme Court Act 1981. An amount of money recoverable as restitution falls within the phrase "a debt or damages" within the meaning of the section, which bites on that amount:

"But section 35A says nothing about the principles to be applied by the courts at the anterior stage when assessing the amount of money required to achieve full restitution."

352. In the next section of his opinion, headed "Measuring the value of the use of money", Lord Nicholls made it clear that the normal measure of such value is an objective one, which has nothing to do with the benefit which the defendant actually derived from the use of the money. He pointed out, at [116], that a comparable objective measure is well established in the law of restitution for wrongdoing, when the court is valuing the benefit derived by a defendant from the unauthorised use of the claimant's land or goods. Even if the unauthorised use does not cause actual damage, "the defendant must still recompense the plaintiff for the benefit he unjustly received".
353. Lord Nicholls continued:

"117. The time value of money, measured objectively in this way, is to be distinguished from the value of the benefits a defendant actually derived from the use of the money. The latter value is not in point in the present case. Sempra retained no proprietary interest in the money it paid to the Inland Revenue, and it has no interest in the "fruits" of that money. Sempra's claim is a personal claim against the Inland Revenue in respect of the benefits it transferred to the revenue. The value of those benefits should be measured as described above.

118. In the present case there can be nothing unjust in requiring the Inland Revenue to pay compound interest, by way of restitution, on the huge interest free loan constituted by Sempra's payment of ACT."

354. Lord Hope similarly approached the question from first principles. Having reviewed Sempra's restitutionary claims, and the development of the law of restitution by Lord Goff of Chieveley in the Kleinwort Benson case, Lord Hope said in [26]:

"[Lord Goff's] speech in the Kleinwort Benson case provides us with another vital building block. Recognition that the court has jurisdiction to award compound interest at common law is a short, but logical, step in the further development of the restitutionary remedy. It follows from the fact that the right to recover money paid under a mistake is available at common law. To treat the choice of remedy in unjust enrichment as discretionary would, in my opinion, be inconsistent with the common law right that gives rise to it."

355. Moving on to discuss the basis of the restitutionary award, Lord Hope emphasised the subtractive nature of the remedy ([31]), and distinguished it from both a claim for an account of profits, and a proprietary claim ([32]). Sempra's claim was instead a purely personal one. Lord Hope continued:

"But, as in cases of property other than money where the claim includes restitution for the value of the use of the asset that was transferred, subtraction of the enrichment from the defendant includes more than the return of the money that was transferred at its nominal or face value. That value, in this case, has already been accounted for. The subject matter of Sempra's claim is the time value of the enrichment. This is the amount that has to be assessed.

33. In this case the enrichment consists, not of the payment of a sum of money as such, but of its payment prematurely. As Professor Birks pointed out, the availability of money to use is not unequivocally enriching in the same degree as the receipt of money: Unjust Enrichment, p 53. But money has a value, and in my opinion the measure of the right to subtraction of the enrichment that resulted from its receipt does not depend on proof by Sempra of what the revenue actually did with it. It was the opportunity to turn the money to account during the period of the enrichment that past from Sempra to the revenue. This is the benefit which the defendant is presumed to have derived from money in its hands, as Lord Walker puts it in para 180. The revenue accepts that the money it received prematurely had a value, but it says that the restitutionary award should take the form of simple interest. I do not think that such an award would be consistent with principle."

356. In my judgment it is clear from this passage that Lord Hope, like Lord Nicholls, saw the normal measure of the use value of money as an objective one, which did not depend on what the Revenue actually did with the money. It was in this context, too, that he dismissed the Revenue's argument that the use value of the money should take the form of simple interest. After the two sentences with which Professor Kay expressed his agreement (see paragraph 349 above), Lord Hope continued:

"Restitution requires that the entirety of the time value of the money that was paid prematurely be transferred back to Sempra by the revenue.

34. All this points to the conclusion, subject to what I say later about onus (see paras 47, 48) that, for restitution to be given for the time value of the money which was paid prematurely, the principal sum to be awarded in this case should be calculated on the basis of compound interest."

357. So far, there is a broad consensus between Lord Hope and Lord Nicholls which I would sum up in the following propositions:

(a) There is jurisdiction at common law to make an award which reflects the time value of the defendant's unjust enrichment by a sum of money which the defendant has received at the expense of the claimant.

(b) Such an award is not discretionary, because the time value of the use of the money is part of the unjust enrichment itself. Restitution will only be achieved if the whole of the undue enrichment, including the use value of the money, is subtracted from the defendant and restored to the claimant.

(c) Nor does such an award depend on proof of what the defendant actually did with the money. Rather, it reflects the opportunity which the defendant had to turn the money to account during the period of the enrichment.

(d) The measure of the time value of the money is therefore normally an objective one, as in cases of restitution for the unauthorised use of a claimant's land or goods.

(e) Such an objective measure requires an award of compound interest, because that is how money is obtained and turned to account in the real world.

(f) An award of this type needs to be distinguished from (i) an account of profits actually made by the defendant from use of the money, and (ii) a proprietary claim to recover the money or its traceable proceeds.

358. Furthermore, I do not believe that there is anything in the above summary with which Lord Walker, the third member of the majority in Sempra, would have disagreed. Lord Walker began his opinion, at [154], by saying that he had read in draft the opinions of Lord Nicholls and Lord Hope, and was "essentially in agreement with them", and "would dismiss [the] appeal, largely for the reasons which they give". Later on, after an illuminating review of the previous law on the topic of interest in unjust enrichment, Lord Walker said at [178]:

"The crucial insight in the speeches of Lord Nicholls and Lord Hope is, if I may respectfully say so, the recognition that what Lord Nicholls calls income benefits are more accurately characterised as an integral part of the overall benefit obtained by a defendant who is unjustly enriched. Full restitution requires the whole benefit to be recouped by the enriched party: otherwise "the unravelling would be partial only" … "

359. In [180], Lord Walker emphasised the need for accurate and consistent terminology, with a view to distinguishing between "(1) proprietary claims which may involve tracing in equity … (2) personal claims for an account of profits (that is, for a sum equal to the profits actually made by the defendant); and (3) personal claims for interest which represents (in a more or less conventional way) the benefit which the defendant is presumed to have derived from money in his hands". Finally, speaking of the judgment of the ECJ in the Hoechst case, Lord Walker said at [183]:

"The judgment of the Court of Justice is in my opinion a powerful encouragement for this House to reconsider the basis on which a monetary award reversing unjust enrichment can and should take account of the time value of money. In modern economic conditions simple interest does not provide full compensation in a case where unjust enrichment has lasted for a significant period …"

360. It is true that Lord Walker went on to say that his preferred solution to the problem would have been to develop and extend the court's equitable jurisdiction to award compound interest, and said he felt "some apprehension" about the suggested conclusion of Lord Nicholls and Lord Hope that compound interest should be available as of right at common law, subject only to an exception to allow for "subjective devaluation" in order to avoid injustice in hard cases: see [184] to [187]. He concluded, however, in [188], that in the instant case either the common law route or the equitable route led to the same conclusion, and the appropriate exercise of discretion (following the equitable route) would be to order the Revenue to pay compound interest at a conventional rate calculated by reference to the average cost of Government borrowing during the relevant period.
361. In Portfolio Dividends (No. 2) I expressed the view, at [239], that in making these observations Lord Walker was not intending positively to dissent from the "as of right" solution propounded by Lord Hope and Lord Nicholls, at any rate in relation to restitutionary claims against the Revenue. I said:

"It seems to me that, in the type of case under consideration, Lord Walker saw either route as providing an acceptable route to a solution (compare the opening words of paragraph [188]), although his preference was for the equitable route. Had he intended to push his disagreement on this critical issue to the point of actual dissent, I think he would have said so explicitly, and would also have expressed his preferred solution in less tentative terms … To summarise, I regard Lord Walker's observations in paragraphs [184] to [188] as being broadly supplemental in character, and not as intended to detract from his basic agreement with the reasoning and analysis of Lords Hope and Nicholls."

With the benefit of the further argument which I have heard in the present case, from different teams of counsel on both sides than in Portfolio Dividends (No. 2), I see no reason to depart from that view.

362. It is also important to note that more recently, in Benedetti v Sawiris [2013] UKSC 50, [2013] 3 WLR 351, ("Benedetti") the Supreme Court has endorsed the proposition that, in valuing the benefit conferred on the defendant, the starting point, or general test, or prima facie position is that the court should adopt an objective approach and seek to determine the market value of the benefit: see the judgments of Lord Clarke of Stone-cum-Ebony (with whom Lord Kerr of Tonaghmore and Lord Wilson JJSC agreed) at [12] to [16], Lord Reed JSC at [99] to [103] and Lord Neuberger of Abbotsbury PSC at [180] to [184]. The question arose in Benedetti in the context of a claim by Mr Benedetti to a quantum meruit for his services in a complex commercial transaction, but all members of the court began by looking at it as a matter of general principle. Moreover, they drew on, and to my mind clearly accepted, the general thrust of the reasoning of Lord Nicholls and Lord Hope in Sempra: see in particular [16], [103] and [182]. As Lord Neuberger said at [182]:

"Further, although the issue involved can be said to be slightly different, namely payment under a mistake, the approach of Lord Hope and Lord Nicholls, in the House of Lords decision in [Sempra] … seems to me, as it did to Etherton LJ at [2010] EWCA Civ 1427 at [144], to indicate that market value is the prima facie basis of valuation in this area of law."

363. The normal market value measure may be departed from in cases of what is sometimes called, rather misleadingly in the view of some judges and academic writers, "subjective devaluation". Lord Nicholls described this principle in Sempra at [119]:

"119. Here, as elsewhere, the law of restitution is sufficiently flexible to achieve a just result. To avoid what would otherwise be an unjust outcome the court can, in an appropriate case, depart from the market value approach when assessing the time value of money or, indeed, when assessing the value of any other benefit gained by a defendant. What is ultimately important in restitution is whether, and to what extent, the particular defendant has been benefited: see Burrows, The Law of Restitution, 2^nd ed (2002), p 18. A benefit is not always worth its market value to a particular defendant. When it is not it may be unjust to treat the defendant as having received a benefit possessing the value it has to others. In Professor Birks's language, a benefit received by a defendant may sometimes be subject to "subjective devaluation": An Introduction to the Law of Restitution (1985), p 413."

364. Lord Hope also dealt with the question at [47] to [48], starting from the (dissenting) view of Lord Mance that the basic test of recovery should be the actual benefit to the defendant. From that perspective, the question could be regarded as one which went mainly, or even entirely, to the onus of proof, and Lord Hope appears at first sight to have accepted this when he said at [48]:

"Once the claimant has shown that prima facie he is entitled to a restitutionary remedy, direct knowledge of the extent of the benefit, if any, that has been received can be assumed to lie with the recipient. It is open to the recipient to demonstrate that there was no actual enrichment when the money fell into his hands notwithstanding the opportunity to turn it to account."

He then said that the Revenue's case was "not that it did not use the money at all", but rather that in view of the financial relationship between the Government and the Bank of England, the benefit was extremely difficult to quantify. On that evidence, Lord Hope considered that the "assumption" that the Revenue derived some benefit from the receipt of the money prematurely had not been displaced, and that this justified resort to a conventional rate of interest as the measure of that benefit.

365. This train of thought was picked up by Lord Clarke in Benedetti, who said at [18] that in his opinion "it is permissible to reduce the objective market value in order to reflect the subjective value of the services to the defendant". He said (ibid.):

"A defendant, in my view, is entitled to prove that he valued the relevant services (or goods) provided by the claimant at less than the market value. That principle is widely accepted by academic commentators and is based on the fundamental need to protect a defendant's autonomy. It is important to note that subjective devaluation is not about the defendant's intentions or expectations but is an ex post facto analysis of the subjective value of the services to the defendant at the relevant time."

He then quoted from Goff & Jones, The Law of Unjust Enrichment, 8^th ed (2011) at para 4-06, including the following passage:

"The common law "places a premium on the right to choose how to spend one's money" … and this right might be unfairly compromised if a defendant were forced to make restitution of the market value of a benefit which he would only have bought for himself at a lower price, or which he would not have bought at all. To avoid this, the court may therefore assess the value of the benefit by reference to the defendant's personal value system rather than the market."

366. After referring to further academic opinion, Lord Clarke continued at [21]:

"21. After the claimant has adduced evidence of the objective value of the benefit which the defendant received, the burden of proof falls upon the defendant to prove that he did not subjectively value the benefit at all, or that he valued it at less than the market price: Goff & Jones, para 4-08; Virgo, pp 64, 66-67. That principle was established by the majority of the House of Lords in Sempra Metals [2008] AC 561: see para 48, per Lord Hope of Craighead, para 116, per Lord Nicholls of Birkenhead and para 180, per Lord Walker of Gestingthorpe. The minority took a different view, namely that it was for the claimant to establish the actual benefit obtained by the defendant: see especially per Lord Mance, at paras 231-232, and Lord Scott of Foscote, at para 147. As I see it, the difference between them is really no more than a different approach to the burden of proof. In each case the question is what was the value to the defendant."

367. Although the principle of subjective devaluation was accepted by the majority for whom Lord Clarke spoke in Benedetti, Lord Reed's preference would have been to deal with the problem by a more nuanced approach to the ascertainment of market value, recognising that the objective value of a benefit to the defendant may be less than its ordinary market value: see his illuminating discussion at [100] to [119]. In Lord Reed's view, the conventional measure adopted in Sempra was an example of a special market value of this type. As he said, at [107]:

"The defendant, as a public body, could purchase the benefit in question (the use of money) at a lower price than commercial enterprises. The benefit arising from the mistaken payment of tax before it was due was therefore valued on the basis of the public sector borrowing rate rather than ordinary market rates of interest."

368. It seems that Lord Clarke would have been prepared to accept this as an alternative way of looking at the matter, while recognising that it made little if any practical difference to the outcome. He said at [22]:

"When I first drafted this judgment I thought that Sempra was an example of subjective devaluation in practice. It was held that the claimant could not recover the market interest rate on the sums it had paid to the Revenue by way of unlawfully levied advance corporation tax because the Government was able to borrow money at lower rates than the market rate. The amount saved by the Government was thus less than that which would have been saved by a commercial entity borrowing the same sums of money: see Goff & Jones at para 4-07. However, having read Lord Reed JSC's judgment I can now see that it may be an example of the objective value of the money to a person in the position of the defendant, namely the Government. This perhaps shows the narrowness of the difference between our two approaches."

Lord Clarke then cited from [119] of Lord Nicholls' speech in Sempra, describing it as "an important passage".

369. It is clear, I think, that the law on so-called subjective devaluation is still at an early stage of development, even though the existence of the principle has now been recognised by the majority in Benedetti. Indeed, Lord Reed said as much at [119]; and see too the relatively agnostic approach of Lord Neuberger at [185] to [192]. At least in the context of claims to recover money paid by mistake, however, it is in my judgment important not to lose sight of the objective nature of the normal measure of benefit, and the related principle that it is generally irrelevant to investigate what the recipient actually did with the money. The market value of the use of money is normally measured by compound interest, and justice will normally be done only by treating the defendant as if he had received a loan of the money on ordinary commercial terms for a borrower in his objective position. That, as I understand it, is the starting point; and in the ordinary case it should also be the end of the enquiry. The normal measure can be displaced, in my view, only if the defendant is able to establish, the onus being on him, that the receipt of the money either did not have its ordinary commercial use value for him, or that it had no use value for him at all (because, for example, he would never have borrowed it in the first place, and after its receipt he merely placed it in a non-interest bearing account until it was repaid). In such a case, as it is put in Goff & Jones at para 4-06, the value of the benefit to the defendant is assessed by reference to his personal value system rather than the market.
370. Even at this second, essentially subjective, stage of the enquiry, it is in my judgment crucial to appreciate that the question is what was the value of the use of the money to the defendant. The question is not what benefit the defendant actually obtained by the use of the money, although the way in which the money was used may be evidence which throws light on the use value of the money to him, i.e. according to his personal value system. To substitute a test, at this stage, of what benefit the defendant actually obtained from the use of the money would in my view be subversive of the normal objective measure, and would in effect reintroduce the approach to valuation which was so powerfully advocated by the minority in Sempra, but (if my analysis is correct) was rejected by the majority in Sempra, as well as by all the members of the Supreme Court in Benedetti.
371. It seems to me that there is a real difference of substance between the approaches of the majority and the minority in Sempra, and that it goes well beyond the burden of proof, important though that is. The critical distinction, to my mind, lies in the need for the defendant to establish a personal value system which means that the money did not, in his hands, have its usual commercial use value. If this can be established by the defendant, the normal objective measure of benefit will be displaced; but otherwise it will prevail, even if the objective measure exceeds the actual benefit obtained by the defendant from the use of the money.
372. Another way of making essentially the same point is to say that what has to be valued, in the normal way, is the defendant's opportunity to use the money. The market value of the opportunity will be the same, once the appropriate market has been identified, whatever the use to which the money is actually put by the recipient.
373. I acknowledge that there is a certain tension between the observations of Lord Hope on the burden of proof, in [47] and [48] of Sempra, and his earlier endorsement of the objective approach to ascertainment of the benefit. A similar tension exists, if I may respectfully say so, in the views expressed by Lord Clarke in Benedetti at [21]. I do not find this surprising, when the subject is one of such difficulty, and the courts are still working their way towards a principled and satisfactory solution. I remind myself, too, that individual judgments, even of the highest court, should not be read as if they were statutes. In my view the observations of Lord Hope and Lord Clarke on the burden of proof do not form part of the ratio of either case, and although entitled to the greatest respect, they are not binding on me. I think I am therefore free to conclude, as I have done, that the difference between the majority and minority in Sempra goes beyond the burden of proof, and that the doctrine of subjective devaluation does not permit the defendant, in every case, to substitute his actual benefit from the use of the money (assuming he can establish it) for its market value as the measure of his unjust enrichment.
374. I have been labouring these points because they are in my judgment of central importance to the issue of quantum in the present case. If subjective devaluation is to be understood in the way I have suggested, it is hard to see how there could be any scope for its operation in the present case (except to the extent that measurement of the benefit by reference to Government rates and terms of borrowing may be thought by some to be an example of the doctrine in practice, rather than an illustration of the objective value of the overpaid sums to the Government in a special market). It could not be argued, with any degree of plausibility, that money is somehow less valuable to the Revenue, or to the Government, than it is to other commercial borrowers. Money may cost the Government less to borrow, but its value is the same, and is reflected in all the different forms which Government expenditure may take. If, on the other hand, the difference between the majority and the minority in Sempra is merely a matter of the burden of proof, it would in principle be open to the Revenue to establish by expert evidence that the actual benefit obtained by the Government from the use of the overpaid tax was exhausted at an early stage, or at least that it would be properly reflected in the payment to the claimants of simple rather than compound interest.
375. Before expressing a concluded view on this question, I will consider what assistance I can gain from the expert evidence of Professor Kay and Dr Richardson, as well as from the views expressed by Vos J in Littlewoods (No. 1).

(2) The expert evidence of Professor Kay and Dr Richardson

(a) Government borrowing costs

376. As I have explained in the context of the adequate indemnity issue, Professor Kay's evidence about the cost to the Government of borrowing the amounts in question was substantially unchallenged, and I accept both his methodology and his calculations without adjustment. Dr Richardson said that the difference between his approach and Professor Kay's was "not fundamental", and he provided no alternative calculations of his own. His suggested modifications of Professor Kay's approach were, first, to use ten-year gilts alone, instead of Professor Kay's five pots, as being representative of Government borrowing; and secondly, to use a moving rather than a weighted average. The former modification, however, would as both experts agreed produce a significantly higher outcome than Professor Kay's more calibrated approach; while Dr Richardson also accepted that use of a weighted average would probably be more accurate than a simple moving average.

(b) The time value of money

377. In his written and oral evidence, Professor Kay lucidly expounded the standard economic theory of the time value of money. The theory reflects the simple truth that the value of money depends on the date when it is received. The difference in value between £1 received in 2014 and £1 received in (say) 1980 is a function of two factors: inflation, and time preference (people generally prefer to have the use of money at once rather than at a future date). In his second report, Professor Kay illustrated the point with a vivid example:

"If I received £1 in 1975, and spent the £1 on four pints of beer, the pleasure that I derive as a result of that expenditure occurs in 1975. I do not derive any benefit in 2005, or in any subsequent year, from that 1975 receipt and the consequential expenditure. However I derive considerable benefit from the fact that I received the £1 in 1975 rather than 2005. There are two reasons for this. The £1, which would have bought four pints of beer in 1975 would not have bought one pint of beer in 2005. And even if the price and availability of beer is unchanged, present consumption may be valued more highly than future consumption. That difference in the value of money received at materially different times (the time value of money) is the quantum of benefit which falls to be measured in the present case. The time value of money is the result of inflation (beer tends to cost more at a later date) and time preference (people prefer to drink beer sooner rather than later)."

378. This example also shows that the time value of money has nothing to do with how it is actually spent or applied. The beer in the example was all drunk in 1975, but it would make no difference if the £1 had been spent on acquiring an asset of durable value, or put in a savings account, or spent on a winning (or losing) bet on the Grand National, or lost down the back of a sofa. An empirical enquiry into how the money was used would be necessary if the relevant question were whether the recipient still enjoyed any continuing benefit from it, or if he were an accounting party who had to account for his stewardship of it, or if a proprietary remedy were sought against him. But those are not the relevant questions. The time value of money is an objective measure of the value to be attributed in 2014 to the receipt of money at an earlier date.
379. The methodology used by Professor Kay in calculating the present time value of the payments of tax is the standard one known as discounted cash flow, which converts all cash flows over an extended period to a common standard by applying compound interest from the actual date of payment or receipt to the present. Professor Kay agreed in cross-examination that the effect of applying this technique is the same as if the Government were treated as having been lent the sums in question at equivalent rates of interest.
380. Professor Kay was cross-examined by Mr Swift at some length about the appropriateness in the present context of using the time value of money as a measure of the benefit to the Government derived from the overpayments, but in my view Mr Swift made little headway and Professor Kay's evidence remained essentially unshaken. Mr Swift's lines of questioning sought in various ways to look at the actual use of the money, or the continuing benefit derived from it by the Government, but I agree with Professor Kay that for present purposes these are the wrong questions to ask. The time value of money is assessed separately from the use of the money, and there is a category confusion if the two questions are conflated. Mr Swift also put it to Professor Kay that the time value of money is distinct from inflation. Professor Kay agreed, and accepted that it is possible to ask simply whether the purchasing power of a sum in the past was the same as it is today. Again, however, that is not the same question as the difference in value between money received on the earlier date and money received today. The time value of money reflects not only the impact of inflation, but also the impact of time preference. If inflation is left out of account, the underlying real rate of interest employed in the discounted cash flow calculation may be seen as the measure of time preference.
381. Dr Richardson's primary thesis was that it is wrong in principle to use the time value of money as the measure of the Government's enrichment, and that the focus should instead be exclusively on the benefit which the Government actually obtained from the use of the money. Even if that approach is wrong, he also maintained that the time value of money is a theory which applies only to the assessment of the present value of future benefits, and cannot be used to quantify past benefits. I found this a rather strange contention, and I have no hesitation in rejecting it. I accept Professor Kay's evidence that the technique of discounted cash flow can be used retrospectively to ascertain the present value of a past receipt of money, even though it is in practice more often used prospectively in the evaluation of future projects. Professor Kay accepted that the extracts from text books attached to his second report all referred to the concept of use value in the context of determining the present value of future sums, and that none of them dealt with the question of ascertaining the present value of a sum received in the past. Professor Kay's explanation for this was that the question arises only infrequently in a retrospective context, but the principle is exactly the same, and it had never occurred to him to draw the distinction suggested by Dr Richardson. He gave the example of a person who invested money in 1973 in order to provide a pension for himself in 2013. A discounted cash flow calculation would have been employed in deciding how much to invest. If the same question were asked in retrospect in 2013, namely how much should have been invested in 1973 to produce a pension of a given amount in 2013, the approach would be exactly the same. He said (day 4, page 98) that the process of discounting "is equally applicable, whether you are looking back or whether you are looking forward". Professor Kay also gave an example derived from his own experience, where (he said) he and some colleagues had written a published article (not in evidence before me) which applied a retrospective cost benefit analysis to the Eurotunnel project.
382. It seems to me that the only distinction which could sensibly be drawn between the prospective and retrospective application of the principle is that future expenditure is by definition uncertain, in the sense that it has not yet happened, whereas past expenditure is a known quantity, and it is possible to see with hindsight exactly how the money was used. But that distinction is irrelevant, precisely because the time value of money is an objective measure which has nothing to do with the actual use of the money. Dr Richardson accepted in cross-examination – he could hardly have done otherwise – that account has to be taken of inflation when comparing an earlier receipt with a later receipt of the same nominal amount. His denial of the retrospective time value of money therefore seems to come down to the contention that time preference is not a meaningful concept when one is looking back at what has actually happened. The fallacy in this reasoning is in my judgment the same as before: it confuses the objective value of the money at the date of receipt with the question of how it was subsequently used.
383. Mr Rabinowitz exposed the unrealistic nature of Dr Richardson's stance on this point with an example which he put to him of two bridges, each of which cost £10 million to build in 2013 and was to be decommissioned in 2020. The first bridge began to yield income in 2015, in an amount which increased from year to year, producing a total in nominal terms of £450,000 by the end of 2020. The second bridge produced no revenue at all until 2020, but in that year it produced £450,000. The total revenue derived from each bridge, in nominal terms, was therefore the same, but the first bridge yielded the £450,000 over a period of years which began in 2015. Dr Richardson's view was that, on a retrospective appraisal of each bridge in 2020, the benefit received from them was the same. He would take no account of the time value of the earlier receipts in the case of the first bridge, as the following exchange (day 4, page 155) makes clear:

"Q. You take no account at all of the time value of getting £50,000 in 2015 in that analysis.

A. If I am asked in 2020, what is the benefit that I have received from the bridge, then, yes, that is correct."

With all respect to Dr Richardson, I do not see how that answer can possibly be correct. The income stream was clearly more valuable to the owner of the first bridge, because he received the money earlier.

384. As to Dr Richardson's primary focus on the actual use which the Government made of the overpaid tax, it seems to me to face two insuperable problems. First, it is contrary to the principles laid down by the majority of the House of Lords in Sempra, and endorsed by the Supreme Court in Benedetti. Secondly, it is entirely irrelevant to ascertainment of the time value of the overpayments. The use made of the money would in principle be relevant to the defence of change of position, which Vos J has rejected. It would also be relevant to any defence of subjective devaluation, but I have already explained why I can see no scope for such a defence in the present case. Apart from those possibilities, however, there is in my judgment nothing to displace the time value of money, measured by reference to the Government's borrowing rates, as the appropriate tool for the court to adopt in ascertaining the benefit to the Government from the overpayments.

(c) The actual benefit to the Government from the use of the money

385. If I am right so far, this is of course the wrong question to ask; but I will nevertheless examine the expert evidence directed to it, in case I am mistaken in my analysis of the law, or if for any other reason a higher court considers that the benefit to the Government should not be measured by reference to the time value of the money which it received.
386. It is common ground that the starting point must be the findings of fact already made by Vos J, in the context of the change of position defence, that the overpayments were not material to Government spending decisions in each fiscal year, and at the end of each year Government borrowing was reduced by the amount of the overpayments in that year: see Littlewoods (No. 1) at [121] to [123]. It would seem to follow from this that the Government will have saved, at least, the interest which it would otherwise have had to pay in the second year on the amount of the original overpayment, because the Government's overall borrowing was correspondingly reduced in that year. Moreover, interest on the original overpayment would on the face of it continue to be saved in the same way, in the third and subsequent years, until the principal amount was eventually repaid to the claimants. But what about the interest on the saved interest? At first sight, it might be thought that, since the original overpayment was too insignificant in size to affect Government spending decisions in the year of its receipt, the same must be true of the much smaller amount of saved interest in year two, with the result that the saved interest must itself have gone further to reduce Government borrowing. If this were indeed the correct model, the saving to the Government from the original overpayment would compound from year to year, and the actual benefit obtained by the Government from the overpayment would appear to equate with compound interest, at the Government's borrowing rates, on the amount of the overpayment; or, in other words, the actual benefit would in fact be the same as the objective measure for which the claimants contend.
387. In broad terms, this was indeed the argument advanced by Professor Kay, on the assumption that his primary approach was incorrect. Dr Richardson, however, argued for a different approach. He accepted, of course, that the amounts of saved interest would not in themselves have been material to Government spending decisions had they been an unknown quantity like the initial overpayments. But, he said, the position in the second and subsequent years was fundamentally different, because the level of Government borrowing at the start of each fiscal year is a known amount which is precisely computed. The interest saved as a result of the overpayment in year one would therefore have been fully factored into the debt interest forecasts for year two and subsequent years, and hence into the public finance forecasts on the basis of which decisions on borrowing, taxation and spending were made.
388. The next stage in Dr Richardson's analysis is that the interest saved did not, either deliberately or by default, go to further reduce Government borrowing. No deliberate decision to that effect was taken, because there was nothing to distinguish the interest saved (or indeed the original overpayments) from the myriad other receipts of comparable size which went into the Consolidated Fund. The Government did not know that it would one day have to repay the overpaid tax, and there was never any question of placing the overpayments in a hypothecated fund. The interest saved therefore formed part of the undifferentiated general pool of resources available to the Government in the second and subsequent years. Moreover, says Dr Richardson, Government borrowing has been tightly controlled throughout the period under review, and the borrowing targets for each year were set by reference to macro-economic considerations such as the rate of inflation, the need for sound public finances, and the need to finance Government debt at acceptable rates of interest. In short, the primacy of the borrowing level was a persistent feature throughout the years in question, and decisions on public expenditure and taxation always had to be tailored to fit the overall limits on Government borrowing. Accordingly, while the balance between spending and taxation would vary from year to year, the public borrowing levels set each year as part of the Budget judgement formed a constant element in the public finances by reference to which taxation and spending plans had to be made. Since the borrowing levels were fixed by reference to macro-economic factors, uninfluenced by the saved interest, it must follow, according to Dr Richardson, that the saved interest did not itself go to reduce public borrowing, but was instead used either to increase spending or reduce taxation.
389. In his second report, Dr Richardson sought to explain why the benefit of the overpayments to the Government did not compound, as follows:

"17. Taking my approach of considering the actual use of the interest avoided, the question of whether the benefit to Government compounded over time hinges on whether the interest avoided was itself saved or whether it was spent.

…

19. In order for the benefit to Government to have compounded in the way that the claimants assert, the Government would have had to have saved the interest avoided in each and every year, the interest on the interest, the interest on the interest on the interest and so forth.

…

21. There is, in general, no reason why interest payments should automatically remain "in the account" rather than being spent and hence no automatic reason for compounding to occur: whether it does or not depends on choices about the use of the interest (or in this case interest avoided).

22. Compounding would require that the amount of Government borrowing was set lower in each year as a consequence of the interest avoided and indeed that this reduction in borrowing was by an increasing amount in each year. The Government would have to have had an in-built bias towards saving rather than spending. I set out in my first Report that this was not the case.

23. Moreover, at the time they were received, the overpayments were not distinguishable from a very large number of similar payments. The only distinguishing feature that has been put forward was the size of the payments, distinguishing them from larger payments. If the Government had had an in-built bias towards saving the interest avoided on these payments, it would presumably have taken the same approach to all other, indistinguishable, payments. As I set out in my first Report, this has the implication that the Government would today have several trillion pounds worth of financial assets, which it clearly does not. Nor is it feasible that debt would otherwise have been several trillion pounds higher than it is today. It is inconceivable that the UK could have borrowed sums of this magnitude.

24. From considering the actual use of the interest avoided, I conclude therefore that the benefit to Government did not compound. The interest avoided was spent in full, or at least in very large part, rather than saved."

390. Dr Richardson elaborated on this evidence in cross-examination. He explained that, as a matter of accounting, the amount to be borrowed by Government is "the difference between the spending plans and the tax plans" (day 5, page 31). He continued:

"Those sums are reconciled precisely in the plans. … We add them up. So as a matter of accounting, if the borrowing amount is unaffected by the interest avoided, then one or other of the other sums must change as a consequence. Either there is more spending on some item or there is less taxation."

Government spending has to be accounted for to Parliament to the nearest £1,000, so the spreadsheets on which departmental estimates are made operate to the nearest £1,000, and the sums are reconciled to that level. The figure for borrowing, which is unaffected, reconciles to the difference between the tax figure and the aggregate expenditure figure. Thus, if an amount equal to the interest saved is reflected in increased expenditure, it will be material to the actual decisions which are taken, even though it will not form a separately identifiable part of the expenditure budget. A little later, Dr Richardson said (day 5, page 34):

"I think the point that is missing from Professor Kay's analysis is that the total of spending, the big number, if you like, is exactly equal to the sum of all the little numbers. It is the sum of all the pennies. It is not approximately equal. It is exactly equal, the system reconciles."

And again:

"The Year 2 position is that you are setting out your plans. Because this is debt interest avoided, you know precisely what the sum is. It is in the debt interest forecast. You cannot do a kind of by and large debt interest forecast. You actually have to pay people precise sums on precise dates, otherwise the Government would default."

391. Professor Kay disagrees with Dr Richardson's approach to this issue. In essence, Professor Kay extrapolates from his approach to the overpayments themselves, which was accepted by Vos J, and says that there is no reason to treat the interest saved any differently. He put the matter in this way in a report which he prepared in March 2011 for the benefit of the ECJ on the reference in the present case:

"14. The effect of lower borrowing by the Government in any year is to reduce the interest payable on Government debt in all subsequent years … This is the principal benefit to the Government of the lower borrowing, since it routinely renews its maturing debt and has no plans to reduce or repay its overall indebtedness. Over the period since 1973, that indebtedness increased in all but two years.

15. The direct benefit of the Government having had the use of the money, therefore, is the reduced interest which the Government will have had to pay in subsequent years as a result of additional receipts or lower expenditures on interest. Specifically, just as the effect of the overpayments is to reduce Government borrowing in the year in which the overpayment is received, so the effect in subsequent years of the lower interest payments which are consequent on the overpayment (and the consequent reduction in the debt size and interest burden) is to increase the funds available to the Government in these subsequent years, and hence allow it to spend more or borrow less.

16. I have reviewed whether the consequential lower interest payments would have been material in relation to the overall revenue and expenditure of the British Government. I have answered this question using the same method, and the same principles I described above, i.e. I have considered whether the budgetary decisions of the UK Government would have been likely to be affected by the anticipated receipt or non-receipt of these amounts. I have concluded that these amounts would also not be material in this sense. I therefore consider in line with my earlier opinion and the judgment of Mr Justice Vos that the interest savings would also have led to further reductions in Government borrowing, rather than lower taxation or additional expenditure.

17. It is also the case that the further reduction in borrowing that is the result of lower interest payments also has the effect of reducing the interest bill in subsequent years. The Governments saves not only interest, but interest on interest, and thereafter interest on interest on interest. It follows that the benefit to Government from the overpayments compounds over time."

392. In paragraphs 26 to 40 of his first report on quantum dated 5 July 2013, Professor Kay explained why he thought it unlikely as a matter of fact that Government spending would have been higher in any year as a result of previous overpayments by the claimants. On the basis of his interest calculations, which I have accepted, Professor Kay said this:

"36. Taking overpayments and interest savings together, the maximum effect on annual Government finances would have amounted to 0.020% of Government expenditure in 2000/1. This compares with the maximum effect of overpayments alone of 0.010% of Government expenditure in 1980/1 (or 0.011% including overpayments still in dispute). The maximum effect of overpayments including those still disputed is 0.012% in 1981/2 and it was this figure which formed the basis of my earlier opinion.

37. In my view, which Mr Neale and Mr Justice Vos accepted, a figure for overpayments which never exceeded 0.012% of total Government expenditure was never material in relation to overall Government revenues and expenditure, in the sense that expenditure decisions would have been affected by anticipated receipt or non-receipt of these amounts. In my view this conclusion on materiality is not affected by the inclusion of compounded interest, i.e. I would have reached the same conclusion on materiality if the relevant figure had been 0.020% of total expenditure rather than 0.012%.

38. I have also considered the potential effect of accumulated liabilities on Government spending. I have estimated the accumulated liability of the Government to the claimants, including compound interest, on a year by year basis during the period. On the assumptions about the basis of compounding set out at paragraphs 44-50 below, the maximum percentage liability is found in 2001/2, and is 0.29% of Government debt.

39. This figure may be considered to approach the threshold of materiality. Although it remains well within the normal margins of error associated with the forecasting of future debt levels, the percentage is not so small that it would necessarily be eliminated by sensible rounding in the presentation of data. The central criterion of materiality which is relevant for the present purpose, however, is whether the figure is of a magnitude such that Government spending decisions would be likely to have been affected by the belief that the liability existed, or did not exist.

40. In the earlier proceedings Mr Neale and I described the fiscal framework which the Government adopted after 1998. For the present purpose, the key element of that framework was the prescription of a 40% limit on the acceptable ratio of the national debt to national income. At the time at which the accumulated liability was largest in relation to total outstanding debt, the ratio of debt to national income was 34.7% and would, if liability for repayment is included, have been 34.8%. In these circumstances, I think it is unlikely as a matter of fact that Government spending would have been higher in any year as a result of previous overpayments by the claimants. It follows that the benefit to the Government of the overpayments is appropriately measured by reference to their impact on funding costs and it is not necessary or appropriate to enquire further into Government expenditure decisions over the period."

393. Commenting on Dr Richardson's approach, Professor Kay said in paragraph 4.8 of his report in reply dated 6 September 2013:

"Dr Richardson appears to envisage a rather different process in which Government continuously adjusts its expenditure in the light of its actual receipts of tax revenue to achieve a predetermined borrowing target. I do not believe this is a realistic account of the Government's budgeting process."

394. Professor Kay maintained these views under cross-examination. Because the amounts of interest avoided were immaterial, Government plans for spending, taxation and borrowing would have been made independently of whether the interest was avoided or not. As with the overpayments themselves, the probability is that the interest avoided would have resulted in lower Government borrowing. The reason for this, in Professor Kay's opinion, is that borrowing is the residual item in any budgeting process, including that of the Government. He illustrated this by an example (day 4, page 55):

"In Year 1 the Government receives, and let us suppose it is unexpected, £10 million which it did not expect to receive from Littlewoods. In Year 2 there is an interest saving which will be less than £1 million in most years as a result of it having received that amount of money in Year 1. The question then becomes whether if Dr Richardson were to go to the Chancellor of the Exchequer and say, we have realised that revenue was going to be £800,000 more next year than we had previously thought it was going to be, the Chancellor would say in that case I shall reconsider my spending plans. My belief is that is not the case."

395. In considering these rival expert opinions, it seems to me that a critical question is whether Dr Richardson is right in his view that Government borrowing limits as set each year in the Budget are, so to speak, a fixed quantity to which spending and taxation decisions in practice have to accommodate themselves, or whether Professor Kay is right to say that borrowing is the default category to which unexpected receipts below the threshold of materiality must be attributed. If Dr Richardson is correct on this point, it is clear to me from his evidence that actual spending decisions are taken at a level of "granularity" which is fine enough to satisfy any test of materiality when compared with the amounts of saved interest, although it is of course impossible to link particular spending decisions with particular amounts of saved interest.
396. On this part of the case, I prefer the approach and reasoning of Dr Richardson. On the basis of Vos J's conclusion that the receipts of the overpayments in year 1 went to reduce Government borrowing, I find Dr Richardson's reasons for distinguishing between the position in year 1 and the treatment of the saved interest in second and subsequent years to be cogent, and it seems to me distinctly more likely that the saved interest would have been absorbed in Government spending in year 2 than that it somehow continued to reduce Government borrowing on a compounding basis. The premise of the latter view is that the amounts of saved interest were below the threshold of materiality which might have influenced spending decisions, so by default they must have reduced borrowing. But the premise appears to me, with the greatest respect to Professor Kay, to miss the point. At the time, there was nothing to distinguish the saved interest from any other receipts or credits which went to increase Government resources, and those resources were (in the aggregate) available to fund Government expenditure of all kinds, together with receipts from borrowing and taxation. Expenditure plans were made (as Dr Richardson explained) at a level of granularity which rounded figures to the nearest £1,000, and departmental spending estimates were essentially an aggregate of large numbers of such individual items of projected expenditure. I can therefore see no good reason why the saved interest should not be treated as part of the general pool of resources which was allocated in year 2 to meet Government expenditure.
397. The fallacy in Professor Kay's approach, as it seems to me, is to treat each overpayment, or instalment of saved interest, as if it was an unexpected bonus which in principle called for separate consideration, but was in fact too small to receive it; whereas the reality is that the overpayments and the saved interest were routine receipts indistinguishable in practice from all other taxation receipts of comparable amount.
398. I should also record at this point that, again with the greatest respect, I have some difficulty in understanding the reasoning which led Vos J to conclude (in Littlewoods (No. 1) at [120] to [124]) that the receipt of the overpaid tax itself in year 1 must have gone to reduce Government borrowing at the end of the year. His conclusion was based on Professor Kay's evidence about the immateriality of the receipts, which were equivalent (as Professor Kay vividly put it) to "3 pence per week or £1.50 per year, as compared to a household income of £30,000 per year". If that approach were correct, however, I cannot see why it would not also apply to all other individual taxation receipts which happened to fall below the level of materiality; and since most individual tax receipts would undoubtedly fall below such a level, there would then seem to be no escape from the absurd conclusion that the majority of tax revenues in a particular year must have gone to reduce overall Government borrowing instead of being used to fund Government expenditure.
399. Vos J's conclusion on the position in year 1 is, of course, final and binding on me as res judicata in the context of the present case; and I am acutely aware that in questioning it I may well be displaying only my own ignorance of economics. I have, however, thought it appropriate to give a brief indication of my doubts for two main reasons. First, since Professor Kay's approach to the treatment of the saved interest is in essence an extrapolation of his treatment of the overpaid tax, my rejection of his reasoning in relation to the saved interest must inevitably call into question the soundness of his reasoning in relation to the overpaid tax. Secondly, since it is a virtual certainty that the decisions of Vos J and myself will be appealed, it may be helpful to a higher court to have a brief indication of my thinking on the question.
400. The approach of Dr Richardson's which I have so far been considering is his (preferred) first approach the result of which is that the interest saved was reflected in its entirety in increased expenditure or reduced taxation, and none of it went to reduce Government borrowing. This approach follows from his evidence (in, for example, paragraphs 38 and 39 of his first report) that the borrowing target for each year is set first, on the basis of macro-economic factors, and spending and taxation decisions are taken afterwards, consistently with the Government's chosen levels of borrowing or saving.
401. Professor Kay was inclined to question this evidence. In his view, the practical reality throughout the periods in issue was that the Government made decisions to set targets about expenditure, revenue and borrowing, "and the budget judgment is actually the difficult job of reconciling all of these because it would always like to tax less, it would always like to spend more and it would always like to borrow less" (day 4, page 37). Professor Kay also disagreed that the borrowing target rested only on macro-economic factors of a generic nature; he said that it also relates "to the amount the Government wants to spend, the amount it thinks it is prudent to borrow and the amount it thinks is appropriate to tax" (day 4, page 39).
402. On these points, however, I again prefer the evidence of Dr Richardson. Professor Kay has no first-hand experience of involvement in the Government budgeting process, whereas Dr Richardson has been intimately involved in it. In particular, from September 2008 until July 2012 he was Director, Public Spending at HM Treasury; and for the preceding ten years, after joining the Treasury in 1998, he worked predominantly on public spending issues.
403. I would also comment that Dr Richardson's evidence about how Government borrowing targets are set is not invalidated by the fact that, as he readily accepts, the target is in practice often missed, or by the fact that there is a margin for error built into all Government forecasts. The fact that Government borrowing forecasts were more often than not exceeded in the years under review (to be precise, in 22 years out of 32) is explained by the phenomenon known to economists as "deficit bias", but does not alter the fact that the forecasts had priority when it came to making spending and taxation decisions for each year. Similarly, the fact that over the same period the average error in the official forecast for Government borrowing in the ensuing year was £2 billion (see paragraph 95 of Dr Richardson's first report) has no impact on the actual decision making process. As Dr Richardson put it, "Knowing that there will be errors does not detract from the need to make decisions based on the forecast" (paragraph 97).
404. For the reasons which I have given, I would accept Dr Richardson's first approach. I can therefore deal more briefly with his second approach, which he puts forward as an alternative if his first approach is rejected. The essence of the second approach is to treat Government resources in the Consolidated Fund as a single, indivisible whole which is applied pro rata to achieve the Government's objectives. Ignoring tax cuts for simplicity (on the basis that their effect is identical to current expenditure of a non-capital nature), Dr Richardson focuses on the amounts actually spent on public services, for which precise data are available, and the amount spent on reducing borrowing. In most of the years under review, total Government borrowing actually increased; but Dr Richardson proceeds on the footing that in these years borrowing might have been higher still, and in that sense some resources were used to reduce borrowing from a still higher level. He estimates the proportion of resources so used as equal to the proportion of total borrowing to total resources for the relevant year. In the few years when there was an actual reduction in Government debt, he uses a proportion equal to the ratio of debt repayment to total Government resources.
405. Dr Richardson's first approach, which is broadly equivalent to payment of simple interest on the overpaid sums, yields (according to his revised calculations and up to his chosen calculation date) a total benefit derived by the Government from the overpayments of £338,815,583, which after deduction of the statutory interest already paid leaves a net amount of £70,656,449. His second approach yields figures of £369,787,481 (calculated to the same date) and £101,628,347 respectively. The difference of approximately £31 million is mainly accounted for by the fact that the second approach includes an element of interest which compounds, on the basis that some of the interest avoided was used to reduce borrowing and thus led to further interest avoided.
406. Professor Kay said little about Dr Richardson's second approach in his written evidence, but in cross-examination he criticised it for producing results that were empirically impossible, in the sense that the Government would never in fact have spent the interest saved pro rata to all its then current expenditure. I found this objection unconvincing. The whole point of the second approach is that it rests on reasonable default assumptions, it being impossible (as everybody agrees) to trace what the saved interest was in fact spent on. A reasonable assumption is not always invalidated by the fact that it may be empirically impossible. Thus, to take an obvious example, Government social and spending policy in certain areas may reasonably be based on an estimate of the average number of children per family, even though it is empirically impossible for a family to have anything other than a whole number of children. Professor Kay also objected that the second approach made no allowance for the fact that much Government expenditure is contractual, and therefore could not be increased proportionately if an extra resource were received. Again, however, this objection seems to me to miss the point that the default assumption is meant to apply across the board, and even if individual items of contractual expenditure could not be increased, there would be no practical obstacle to making larger increases elsewhere.
407. To conclude, I regard Dr Richardson's second approach as a reasonable fallback position, and I would adopt it as an appropriate way of measuring the actual benefit to the Government of the overpayments if (a) that were the appropriate question to ask, and (b) Dr Richardson's first approach were to be rejected.

(3) The views of Vos J in Littlewoods (No. 1)

408. I have already referred to the views expressed by Vos J in [120] to [125] of Littlewoods (No. 1), in the context of his rejection of the Revenue's defence of change of position. I have also ventured to suggest some disagreement with his approach (based on Professor Kay's evidence) to the question of materiality, and with his conclusion that the whole of the amounts of the overpayments necessarily went to reduce Government borrowing at the end of the fiscal years of receipt. It would not be appropriate for me in this judgment to attempt in any way to re-open the question of change of position, since it has already been finally determined at first instance. But I should briefly notice an argument in the claimants' written submissions, to the general effect that it is not open to the Revenue to run their present case on quantum, based on Dr Richardson's two alternative approaches, because those approaches are inconsistent with the Revenue's arguments on change of position which have been rejected by Vos J.
409. There are certainly some close similarities between Dr Richardson's approach, especially perhaps in his second report, and the arguments which Vos J has rejected; but I do not believe that there is any necessary inconsistency, or that the Revenue should be precluded from running their case on quantum at first instance.
410. In the first place, quantum was not in issue at the First Trial, and Vos J clearly had this well in mind. Secondly, there is at least one important difference between the change of position defence and quantification of the benefit which the Revenue actually obtained from the overpayments. Expenditure of the sums overpaid, or of saved interest, could only found a defence of change of position if it was expenditure which would not otherwise have been incurred; whereas the question of actual benefit derived from the overpayments is a purely factual enquiry, and the relevant question (given Vos J's conclusion that the overpayments went to reduce government borrowing at the end of year 1) is simply whether the saved interest was itself spent, and if so when.
411. Vos J drew this distinction, in my view correctly, when he considered a similar objection in the context of exhaustion of benefits: see [129] to [130] of his judgment. After holding that it would be open to the Government to establish, if it could, at the quantum hearing that the benefit received by it from the overpayments is not properly represented by an award of compound interest at conventional rates, he said:

"130. I have, however, now to consider whether this could be consistent with the findings I have made under issue 4C above. In other words, is it consistent to say, on the one hand, that there has been no change of position because the overpayments must be taken to have decreased Government borrowing in the fiscal year in question, and on the other hand that it is open to the Commissioners to prove that Government has not benefited in fact from the overpayments after the fiscal year in which they were paid. There is, I think, no inconsistency here. The difference is between a defence of change of position, properly so called, which requires it to be shown that a spending decision has been taken in reliance on the overpayment, and a quantum assessment which enables the recipient to show that it has, in fact, not benefited beyond the capital payment because, in this case, of the special government mechanisms for setting its budgets, its borrowing limits and its expenditure."

412. The separate defence of exhaustion of benefits advanced by the Revenue at the First Trial was an extreme one, because it asserted (as pleaded) that all benefit received by the Government from the overpayments ceased after the fiscal year of receipt, or at any rate by the end of a single budget cycle following receipt of the overpayment: see paragraph 46A of the re-re-amended defence and Littlewoods (No. 1) at [127]. The implication of this defence, had it succeeded, is that restitution of the Government's unjust enrichment would have required only repayment of the principal amounts of the overpayments, together with approximately one year's simple interest. While Vos J rejected the defence at the First Trial in so far as it was properly to be regarded as a defence of change of position, he expressly said in [131] that it would be open to the Revenue to argue on the assessment of quantum "that the Government has not benefited from the money after the first fiscal year, so that no use value of the money should thereafter be awarded to Littlewoods". It should be noted, however, that this is not in fact the argument which the Revenue have now advanced. They now accept, realistically in my view, that the benefit derived by the Government from the overpayments must have continued until they were repaid, and that this benefit would be properly reflected in payment of simple interest for the whole period at the rates proposed by Professor Kay.

(4) Quantification: conclusions on the measure of benefit

413. Having considered the expert evidence, and the views of Vos J in Littlewoods (No. 1), I see no reason to depart from the initial conclusions which I drew from Sempra and Benedetti. In my judgment the benefit to the Government from the overpayments of VAT is, as a matter of law, correctly measured by the objective use value of the money, which translates into an award of compound interest computed in accordance with Professor Kay's rates and methodology. The actual use made by the Government of the overpayments, and the actual benefit which the Government derived from them, are in my opinion irrelevant to the objective measure thus ascertained.
414. If, contrary to my primary conclusion, it is open to the Government to establish and rely on the actual benefit which it derived from the overpayments, I consider that the Government has succeeded in showing on the balance of probabilities that the saved interest on the principal amounts of the overpayments was absorbed in Government expenditure in the fiscal year following the year of receipt of the overpayments, and therefore did not further reduce Government borrowing. This finding would in my view be appropriately reflected by payment of simple interest in accordance with Dr Richardson's first approach. If, therefore, contrary to my primary conclusion, the correct approach to quantification were to identify actual benefit, I would find that the total quantum of the claim up to the date of calculation in Dr Richardson's report was £70,656,449. This finding would, however, be subject to the questions (which I have yet to consider) whether (a) the Revenue's approach to quantification is inconsistent with EU law, and (b) the claimants should give credit for the extra corporation tax which they would have had to pay if the overpayments of VAT had not been made.
415. If Dr Richardson's first approach is for any reason to be rejected, I would (subject to the same proviso) accept his second approach, which appears to me to represent a reasonable fallback position and does not in my judgment suffer from the conceptual flaws alleged by the claimants. On this second approach, the total quantum of the claim up to the date of calculation in Dr Richardson's report subject as aforesaid, would be £101,628,347.
416. It follows that I reject Littlewoods' arguments, supported by parts of Professor Kay's evidence, that the actual benefit to the Government from the overpayments compounded from year to year, on the footing that the interest saved itself went to reduce Government borrowing in the second year, and so on in subsequent years. I do also respectfully question whether Vos J was correct to conclude that the whole of each overpayment necessarily went to reduce Government borrowing at the end of the fiscal year of receipt, although I accept that this point has been finally determined by him at first instance.
417. It is also convenient to deal here with another point which I have not yet separately addressed. On the assumption that my primary conclusion is correct, should compound interest continue to run to the date of judgment, or should it stop running when the relevant repayments of principal sums were made and be replaced at that point by simple interest under section 35A of the Senior Courts Act 1981? Neither side addressed me at any length on this question, because I have recently considered it, and resolved it in the taxpayers' favour, in Portfolio Dividends (No. 2) at [245] to [246]. In short, it seemed to me that although there had been no appeal in Sempra from the decision of Park J that interest should run pursuant to section 35A for the period from utilisation of ACT until judgment, the logic of the majority speeches in the House of Lords showed that compound interest should also be available in respect of the post-utilisation period. I confirm that I remain of the same opinion, and I would therefore answer the comparable question in the present case in Littlewoods' favour. The Revenue did not, of course, concede the point before me, but recognised that it would be more sensible to reserve it for a higher court since I had so recently considered it myself and decided it against them.

(5) Is the Revenue's approach to quantification inconsistent with EU law?

418. I will deal with this question briefly, not because it is unimportant (far from it) but because my answers to it either follow from my views on the adequate indemnity issue (which I have already discussed at some length in this judgment) or, in relation to the additional corporation tax issue, largely depend on the decision of the Grand Chamber of the ECJ in September 2011 in the Lady & Kid case (Case C-398/09, Lady & Kid A/S and others v Skatteministeriet, [2012] STC 854). In Portfolio Dividends (No. 2) I considered the impact of Lady & Kid and subsequent ECJ case law on the Revenue's change of position defence in that case, concluding that EU law prevents reliance on change of position as a defence to San Giorgio claims, whether they are brought in mistake pursuant to the DMG cause of action or in reliance on the Woolwich principle.
419. If I have correctly understood the judgment of the ECJ in the present case, it seems clear to me that the claimants will only receive an "adequate indemnity" for their loss occasioned by the overpayments of VAT if they are paid, at least, a sum which represents the use value of the overpayments in the hands of the Government. Strictly speaking, the use value which EU law entitles them to receive is the lost use value to themselves of the sums overpaid. It is common ground that such value would be greater than the use value to the Government, given the Government's ability to borrow at below ordinary market rates. The claimants are not, of course, compelled to seek to recover the maximum remedy afforded to them by EU law, and nobody suggests that they have forfeited their right to claim compound interest merely because they are content to limit their claim to the use value of the overpaid tax in the Government's hands rather than their own.
420. It was necessary to award compound interest in Sempra so as to satisfy that company's right under EU law to be paid the use value of the prematurely levied ACT. Now the ECJ has confirmed that there is a right to interest under EU law in respect of all repayments of taxes levied contrary to EU law, the same consequence must in my judgment follow. Only compound interest will suffice to satisfy the claimants' EU law right to interest. I would therefore hold that the Revenue's attempt to limit the restitution which it has to make to the benefit which it actually received from the overpayments is contrary to EU law, even if it be assumed that, in a purely domestic context, this is all the English law of restitution would require.
421. The next question is whether it is open to the Revenue under EU law to argue that the restitution awarded to the claimants should be reduced by the additional corporation tax which they would have had to pay if the overpayments of VAT had never been made. In my opinion it is clear from Lady & Kid that this question must be answered in the negative. As I explained in Portfolio Dividends (No. 2) at [172], the background to the case was the introduction by Denmark in 1987 of a business tax known as the employment market contribution, or "AMBI". It was charged at a fixed rate of 2.5%, and calculated on the same basis as VAT. It was payable upon the first sale in Denmark of imported goods. In return for the introduction of the AMBI, a number of social security charges which had to be paid by Danish employers were abolished. The purpose of the combined package of measures, i.e. the introduction of the AMBI and the abolition of the social security charges, was to stimulate growth and promote employment, while retaining neutrality with regard to public finances. The AMBI was levied between 1988 and 1991, but in 1992 a challenge to its lawfulness was upheld by the ECJ. Following that judgment, Denmark implemented legislation laying down arrangements for reimbursement of the AMBI which had been unlawfully levied, but various conditions had to be satisfied by claimants for reimbursement, including that they had suffered a net loss when the amount paid by way of the AMBI was set against the savings from the social charges which had been abolished. In the national proceedings before the Copenhagen District Court, there were four claimant retailers, two of which had made a net gain, in the sense that they had saved more in employer social security contributions than they had paid in respect of the AMBI; the other two had made a net loss, and one of them had made no saving at all. On appeal from the rejection by the District Court of all four claims, a number of questions were referred by the appellate court to the ECJ, including a question asking whether an unlawful levy had to be reimbursed where the taxpayer had in fact made a net saving as a result of the abolition of other charges.
422. Rejecting the advice of Advocate General Cruz Villalón in his opinion, the Grand Chamber held that the sole exception recognised by EU law to the right to reimbursement of tax levied in breach of EU law is the defence of passing on, that is to say where the burden of the unlawfully levied tax has been directly passed on by the taxpayer to a third party. Accordingly, the unlawfully levied AMBI had to be refunded to all four claimants, including the two which had made a net gain from the package of fiscal measures of which the AMBI formed part.
423. The core of the Court's reasoning is contained in the passage which I quoted in Portfolio Dividends (No. 2) at [174]. For present purposes, it is enough to cite the following brief extracts:

"20. None the less, since such a refusal of reimbursement of a tax levied on the sale of goods [i.e. the established passing on exception] is a limitation of a subjective right derived from the legal order of the European Union, it must be interpreted narrowly. Accordingly, the direct passing on to the purchaser of the tax wrongly levied constitutes the sole exception to the right to reimbursement of tax levied in breach of European Union law.

…

25. … It must be noted that the Court, in paragraph 20 of the present judgment, states that the direct passing on of the tax wrongly levied to the purchaser constitutes the sole exception to the right to reimbursement of tax levied in breach of European Union law."

424. In my judgment the stark principle enunciated by the ECJ in Lady & Kid (and followed in subsequent cases such as Accor (Case C-310/09, [2012] STC 438) must now be taken to represent the considered view of the ECJ on the availability of defences to a San Giorgio claim. Furthermore, the facts of the cases referred show that the strictness of the principle cannot be circumvented by allowing a deduction for countervailing fiscal benefits obtained by the claimant in connection with the imposition of the unlawful levy. Since it is now clear that the right to interest on unlawfully levied tax is itself protected by EU law, it seems to me that the same principles must apply in relation to interest, with the result that the claimants in the present case would be deprived of an "adequate indemnity" for their loss if their claims were reduced to take account of the additional corporation tax that would have been paid. As the ECJ said in Lady & Kid at paragraph 26:

"… European Union law precludes a Member State from refusing reimbursement of a tax wrongfully levied on the ground that the amounts wrongly paid by the taxpayer have been set off by a saving made as a result of the concomitant abolition of other levies, since such a set-off cannot be regarded, from the point of view of European Union law, as an unjust enrichment [i.e. of the taxpayer] as regards that tax."

425. In the present case, the alleged benefit to the claimants arose not from the abolition of another tax, but rather from a reduction in the amount of another tax as a result of the payment of the unlawful VAT. In essence, however, the position is very similar: but for the unlawful tax, a different tax would have been imposed on the same taxpayer. I can see no good reason why the same principles should not apply.

(6) Should the saving of additional corporation tax be taken into account under English law, and (if so) how should it be computed?

(a) The position under English law

426. I come finally to the questions whether, as a matter of English law, the saving of additional corporation tax should be taken into account in quantifying the claims, and (if so) how the saving should be computed. These questions do not arise if I am right in my conclusion (see paragraphs 421 to 425 above) that EU law prevents the Revenue from raising them. It is also hard to see how they could arise if I am right in my view that the English law of restitution requires payment to the claimants of the full use value of the overpaid tax, without reference to the actual benefit obtained by the Government from the overpayments. But if it were in principle open to the Revenue to establish actual benefit as the measure of the restitution to be made, and assuming this not to be prohibited by EU law, how would the saving of corporation tax fit into the jigsaw puzzle?
427. It is important to remember at the outset that, even making the above assumptions in the Revenue's favour, the claims remain ones for the restitution of unjust enrichment. It is therefore necessary to find some principle in the English law of unjust enrichment which would allow account to be taken of the additional corporation tax. The claims are not ones for damages, where in quantifying the claimant's loss the court will often take into account countervailing benefits to the claimant arising from the actionable wrong.
428. The claimants submit that for a number of separate reasons the additional corporation tax cannot be taken into account. First, since the additional tax was never in fact levied or paid, they say that to take it into account would offend the constitutional principle that tax cannot be recovered without the authority of Parliament. In substance, submit the claimants, the Revenue are attempting to collect a tax, or interest on a tax, which is not and never has been due. Secondly, there is no authority in the law of unjust enrichment which supports the Revenue's approach on this point, and such authority as there is points firmly in the other direction. Thirdly, the law does recognise a principle of counter-restitution, but it operates within narrow confines and is not applicable on the facts of the present case.
429. I can rapidly dispose of the first of these objections. I agree with Mr Swift that taking account of the corporation tax saved by the claimants would not offend against any constitutional principle. The Revenue are not attempting to levy the tax which has been saved, but simply arguing that any realistic assessment of the actual benefit to the Government from the overpayments of VAT must make due allowance for the hypothetical extra corporation tax which would have been paid by the claimants. No constitutional principle is engaged by the making of counter-factual assumptions.
430. Turning to the authorities, the case which Mr Rabinowitz placed at the forefront of his submissions was the decision of the High Court of Australia in David Securities Pty. Ltd v Commonwealth Bank of Australia [1992] HCA 48, (1992) 175 CLR 353. In my view, however, the facts of David are too far removed from the present case to throw any useful light on the question. If anything, it seems to me that the Revenue can find some support in the judgment of the majority, delivered by Mason CJ, Deane, Toohey, Gaudron and McHugh JJ, which includes the following passage at p 379:

"The fact that the payment has been caused by a mistake is sufficient to give rise to a prima facie obligation on the part of the respondent to make restitution. Before that prima facie liability is displaced, the respondent must point to circumstances which the law recognises would make an order for restitution unjust. There can be no restitution in such circumstances because the law will not provide for recovery except when the enrichment is unjust. It follows that the recipient of a payment, which is sought to be recovered on the ground of unjust enrichment, is entitled to raise by way of answer any matter or circumstance which shows that his or her receipt (or retention ) of the payment is not unjust."

The two particular defences upon which the respondent relied in David were that the payments in question had been made for good consideration, and change of position; but the passage which I have quoted suggests that there are no a priori limits to the kinds of circumstance which may show that the receipt or retention of a payment made by mistake is not unjust. Indeed, it seems to me that it is precisely such considerations which have led the English courts to recognise the principle of subjective devaluation. When I put this point to Mr Rabinowitz, he said he could not argue against it (day 9, page 150).

431. Mr Rabinowitz's next point was that the law will only permit a netting-off of benefits where the benefits arise directly from the same transaction as the mistaken payment, and where they are actual benefits which have been transferred, not hypothetical or future ones. He relied for these propositions on the decisions of Hobhouse J in Kleinwort Benson Ltd v Sandwell Borough Council [1994] 4 All ER 890, and Kleinwort Benson Ltd v South Tyneside Metropolitan Borough Council [1994] 4 All ER 972. Thus in the Sandwell case Hobhouse J treated a succession of void interest swap transactions between the same counterparties as analogous to a single running account, with the successive payments merely altering "the location and extent of the enrichment which existed from time to time": see 929f-h and 940g-941e. By contrast, in the South Tyneside case there were five different interest rate swap contracts between the same parties, and because each was an independent transaction Hobhouse J refused to recognise any right of set-off or aggregation as between one contract and another: see 979h-j.
432. As to counter-restitution, the Revenue go no further than to submit that there is an analogy with the principle, which is usefully discussed in Chapter 31 of Goff & Jones, The Law of Unjust Enrichment, 8^th edition (2011). In his Restatement of the English Law of Unjust Enrichment, Oxford (2012), Professor Burrows formulates the principle as follows:

"(1) The defendant has a defence if it is impossible for there to be counter-restitution of reciprocal benefits conferred on the claimant by the defendant.

(2) The defendant normally has a set-off defence for counter-restitution of reciprocal benefits conferred on the claimant by the defendant."

See too the discussion at pp 128-131. The principle is still at an early stage of development, but as currently understood it does not seem to me to extend beyond cases in which a reciprocal benefit has actually been conferred on the claimant by the defendant, often in circumstances where it might in principle be open to the defendant to make an independent claim in restitution in respect of the benefit conferred. The principle merges into the ordinary law of set-off, which applies to restitutionary claims in the same way as it does to other claims. There is a divergence of opinion whether it should be afforded separate status as a defence, or merely viewed as an aspect of the general defence of set-off: see Burrows, p 130.

433. The Revenue also seek to derive support from some examples given by Professor Burrows in the third edition (2011) of his work on The Law of Restitution, when discussing the concept of "incontrovertible benefit" in the context of the question whether the defendant has been enriched. The payment of money is typically regarded as an incontrovertible benefit to the defendant, but Professor Burrows gives reasons for suggesting that this is too extreme a view, because "the description of "incontrovertible" misleadingly indicates that there can never be a rebuttal" (page 50). In this context, he gives the following examples (ibid.):

"Say, for example, the claimant pays £1,000 to the defendant by mistake. As a result a third party does not pay the defendant the £1,000 that she would otherwise have paid him. Or say the claimant has discharged the defendant's liability to a third party but a relative of the defendant was about to discharge it gratuitously. Or say the claimant has improved the defendant's car, and the defendant has realised those improvements, but the defendant would otherwise have made the same improvements at a cost well below the market price of the services. In all these circumstances it would seem correct to say that the defendant has not been benefited at all, or has not been benefited at the market price, by the claimant's payment or services because it would have been benefited in any event or would otherwise have been benefited at a lower price than the market price."

See too Professor Burrows' Restatement, section 7(3)(a), and the discussion of it at pp 42-43.

434. I do not find this an easy question, because there is an initial attraction in the Revenue's simple point that the claimants would undoubtedly have paid more in the way of corporation tax if the unlawful payments of VAT had never been made. On balance, however, I am inclined to agree with the claimants that it would be wrong to take account of the additional corporation tax which the Government would have received when quantifying the actual benefit to the Government of the overpayments of VAT. My reasons are these:

(1) First, the Revenue never had an actual claim for the additional corporation tax. It was always a hypothetical claim, in a counter-factual world where the VAT had not been overpaid. There can therefore be no question of an actual set-off, or of an actual counter-benefit conferred on the claimants by the Revenue.

(2) Secondly, the liability of a claimant to corporation tax would have depended on the quantum of its taxable profits in the accounting period in which the VAT was overpaid, whereas the supposed liability to VAT was a liability to a tax of a fundamentally different nature on the supply of goods or services, computed and accounted for under an independent and self-contained statutory regime. It cannot therefore be said that the two liabilities to tax arose out of the same underlying transaction, or that there was anything akin to a single running account between the parties.

(3) On the basis of the existing English authorities, the position is not analogous with that considered by Hobhouse J in the Sandwell case, and is by contrast much closer to that considered by him in the South Tyneside case.

(4) The facts do not fall within, or even close to, the principles of either subjective devaluation or counter-restitution as they are currently understood.

(5) If the additional corporation tax were to be netted-off, it could only be on the basis of a general principle of considerable width and uncertainty which may have a place in the quantification of damages for wrongs, but seems to me to be at odds with the essentially subtractive nature of an unjust enrichment claim. The Government was undoubtedly enriched by the overpayments of VAT as and when they were received. The benefit of the payment of that money may not be strictly incontrovertible, but the presumption that the Government was thereby enriched by the full amount of the receipts must be a very strong one. That is the benefit which needs to be restored to the claimants, and the justice of the claim to restitution is not in my opinion affected by the probability that, but for the overpayments, the Government would subsequently have received larger amounts of corporation tax than it did in fact receive.

(6) I do not dismiss the possibility that the corporation tax position may be relevant to the Revenue's defence of change of position; but that defence is not before me, and has already been rejected on the facts by Vos J. (I am, of course, also assuming here that the defence of change of position is not precluded by EU law).

(b) Quantification

435. If, contrary to what I have just concluded, the additional corporation tax should be taken into account, how should it be quantified? I have already said a little about this question in the context of the adequate indemnity issue: see paragraphs 308 to 310 above. In resolving it, I have the assistance of expert evidence from Mr David Prestwich on behalf of the claimants and Mr Christopher Barker on behalf of HMRC. Mr Prestwich is a tax partner in Mazars LLP, with over 15 years' experience of advising large companies and groups in respect of their UK and international corporation tax affairs. Mr Barker is a deputy director of the Large Business Service at HMRC. He qualified as an inspector of taxes in 1993, and since then has worked in a variety of tax compliance and leadership roles for the Revenue. In particular, he has extensive experience of reviewing the corporation tax affairs of large businesses. Both experts were cross-examined at the hearing.
436. My task is simplified because the experts have been able to reach agreement on a wide range of matters. Apart from agreeing the amounts of the VAT overpayments, and their allocation to specific accounting periods, they have also agreed:

(a) that due to the limited availability of information about the tax computations of individual companies and the utilisation of group relief, the most reasonable approach to calculation of the additional corporation tax is by reference to the consolidated group accounts which are publicly available for the material period;

(b) that if the overpayments had not been made, turnover and profit before tax would have increased at group level (all other things being equal); and

(c) that no additional corporation tax would have been paid by the Littlewoods claimants in the years 1981, 1982 and 2005.

They also agree (paragraph 2.1.8 of their joint report dated 27 September 2013) that there is evidence that the claimants used corporation tax management strategies "in accordance with the normal behaviour of large companies", and that in practice the effective tax rate ("ETR") of large companies "nearly always differs from the statutory rate".

437. The sole area of disagreement between the experts concerns the rate of corporation tax that should be applied to the overpayments of VAT in order to calculate what they call "the set-off amount". Mr Barker's view is that the full nominal rate of tax should be applied, except where specific information from the group accounts or tax computations points to a different conclusion, while Mr Prestwich favours the application of an adjusted ETR. Each expert considers that his approach is likely to produce a more realistic result. They also agree that the difference between them "relates to the level of assumptions that need to be made in order to produce a realistic calculation for the Set-Off Amount" (joint report, para 3.1.4).
438. In a little more detail, Mr Barker's methodology was essentially as follows. For six years (1981, 1982, 2001 to 2003 and 2005) he has assumed that the Littlewoods group would have made use of otherwise unused reliefs to offset the additional pre-tax profit, with the result that no additional corporation tax would have been due. It is common ground that in 2005 the Littlewoods group was loss-making, and Mr Barker therefore considers it reasonable to assume that there would have been unused loss relief available to reduce the additional profits to nil. In the other five years, the evidence shows that the Littlewoods group was profitable but did not in fact pay any corporation tax. Since various reliefs must have been used in those years to extinguish the pre-tax profit, Mr Barker is prepared to assume in the claimants' favour that further unused reliefs would have been available in amounts sufficient to cover the additional profits. With the exception of these six years, however, Mr Barker has proceeded on the assumption that every extra £1 of pre-tax profit would have translated into an extra £1 of profit chargeable to corporation tax, with the result that the whole amount of the additional profit would have been chargeable to tax at the full nominal rate. In sum, therefore, the effect of Mr Barker's approach is to assume that in six years no extra corporation tax would have been payable, but in all the other years tax would have been chargeable at the nominal rate on the full amount of the extra profit.
439. Mr Barker sought to justify his methodology by saying that it had the merits of simplicity and the avoidance of speculation. I think it is fair to say that he wished to make as few assumptions as possible, on the basis that this was in principle the most reliable way of producing an accurate estimate of the additional corporation tax. But there is a fallacy in this approach, which Mr Rabinowitz exposed in cross-examination. The exercise upon which the experts were engaged was intrinsically one which required the making of assumptions, about what the companies would in fact have done with the extra money. That is a question to be determined on the balance of probabilities, with only little help to be gained from the surviving tax computations. The experts were only able to examine 25 computations, whereas by Mr Barker's own reckoning over 300 would have been required to get a full picture for the claimants over the relevant period, and potentially thousands for a precise calculation of the impact on the overall corporation tax position at group level, taking account of group relief. In such circumstances, there is no obvious merit in making either the simplest or the fewest assumptions. The question is rather what assumptions it is most reasonable to adopt in answering what is inevitably a hypothetical question, namely what the companies would have done if their pre-tax profits had been correspondingly greater between 1973 and 2005. The issue for me is which approach is the more realistic, and that depends on the plausibility of the assumptions which the experts have made.
440. Despite Mr Barker's distrust of assumptions, there are in fact two which underpin his approach to the years other than the six when he accepts that no additional corporation tax would have been payable. The first assumption is that no unused reliefs would have been available to set against the extra profits. The second assumption is that no steps would have been taken to generate extra reliefs, for example by incurring further capital expenditure which would have qualified for capital allowances. In my judgment, however, neither assumption stands up to scrutiny.
441. As to the first, Mr Barker was constrained to accept in cross-examination that in eight out of the 25 cases where corporation tax computations survive, the company in question had substantial amounts of unclaimed allowances. It is fair to say that the majority of these cases related to companies in the Littlewoods Group during years when Mr Barker accepts that no additional tax would have been payable. But that is not true of all of them. For example, the computation of Littlewoods Home Shopping Limited for the year ended 30 April 2004 showed unclaimed capital allowances of over £2 million. Similarly, the computation of Littlewoods Retail Limited for the year to 30 April 2000 showed unused allowances of over £890,000. These were years when the Littlewoods group as a whole was paying corporation tax, so there is no obvious reason why use could not have been made of these unused allowances, either directly by the company concerned or by way of surrender of group relief.
442. As to the second assumption, Mr Barker himself accepts in paragraph 63 of his report that it would in principle be reasonable to assume, as Mr Prestwich has, that the companies in question would have incurred additional expenditure giving rise to additional relief against tax. Mr Barker refuses, however, to act on that reasonable assumption, because "it would be necessary to speculate as to the amount and tax treatment of any hypothetical additional expenditure". That comment appears to me to miss the point, and to illustrate the basic fallacy in Mr Barker's approach. If the assumption is a reasonable one to make, the question should then be how best to reflect it in the calculations. It is not sensible to reject the assumption merely because giving effect to it is necessarily a hypothetical exercise. To take another example, Mr Barker accepted in cross-examination that there is a possibility that the company in question might have purchased additional stock, thereby (subject to certain qualifications) generating additional stock relief.
443. By contrast, Mr Prestwich's approach concentrates on the effective rate of corporation tax. Drawing on published data and his own extensive experience, he explains in his reports that the average ETR of large companies normally differs significantly from the statutory rate of corporation tax, and that in the case of general retailers "the median ETR has generally been lower than the statutory CT rate" (paragraph 4.4.6 of his first report). It is common ground, as I have already noted, that the claimant companies actively managed their corporation tax exposure in accordance with the normal behaviour of large companies. Mr Prestwich therefore considers it implausible to assume that corporation tax would have been payable at the full nominal rate on the whole of the additional pre-tax profits. As a result of active tax management, the additional profits actually chargeable to corporation tax would normally have been substantially lower than the profits before tax as recorded in the claimants' accounts. Mr Prestwich therefore takes as his starting point the claimants' ETR for each year, calculated on the basis of the consolidated group accounts of the GUS and Littlewoods groups. He then makes certain adjustments to allow for exceptional items, overseas taxation and profits, and various other matters (such as deferred taxation, amortisation of goodwill created before 2002, and adjustments relating to prior accounting periods) which have no impact on the amount of a company's corporation tax liability. Having made the appropriate adjustments, Mr Prestwich then calculates the adjusted ETR to be applied to the additional profits. He performs this exercise for all years, except in the few instances where (at the date of his first report) the Revenue accepted that no additional corporation tax would have been payable.
444. In their written closing submissions on this part of the case, counsel for HMRC make a number of criticisms of Mr Prestwich's approach and submit that Mr Barker's methodology should be preferred. I have already explained why Mr Barker's methodology appears to me to be flawed in principle, and counsels' detailed submissions do not cause me to change my mind. For example, on the issue of unused reliefs it is suggested that no conclusions can safely be drawn from the eight computations where there is evidence of unused reliefs, because (a) five of them relate to years in which the Littlewoods group was not paying corporation tax, so it is not surprising if some reliefs were unused, (b) the other three computations may not have been in final form, and (c) even if there were surplus unused reliefs in those three years, there must have been a specific reason why they were not utilised, and in the absence of evidence to the contrary it is reasonable to assume that they would not have been utilised against the additional profits. I accept that there is some force in the first of these points, but the second and third seem to me essentially speculative. The computations in question were supplied to the Revenue, which has given disclosure of them, and there is no firm evidence to suggest that they were in any way provisional. Nor can it be assumed with any confidence that a company would have chosen to leave the relevant reliefs unutilised even if its profits had been higher, merely because it was content to do so as matters actually stood. Without a synoptic view of the tax affairs of the group as a whole, and the strategies pursued by its tax managers, it is really impossible to form any view on specific questions of this sort.
445. As to Mr Prestwich's methodology, HMRC submit that it is important to have in mind the context in which the additional profits would have arisen. The starting point is that higher margins would have been achieved by not having to account for VAT, but there is no reason to assume that the groups' underlying trading activities would have been any different. For example, neither expert has assumed that there would have been additional sales generating additional turnover and profits before tax. On the contrary, the overpaid VAT reflected the groups' existing turnover, and the relevant assumption is simply that the amounts paid as VAT would have been added back into the accounts. Thus, it is said, the most reasonable assumption to make is that the marginal additional profits would simply have been taxed at the statutory rates.
446. I do not accept this submission, which seems to me to rest on the false premise that the saving in VAT should be likened to an unexpected cash bonus. In reality, the profits of the relevant companies would have been correspondingly higher, but the saving of VAT would not have formed an identifiable separate part of those profits. Given that each group pursued tax management strategies which had the result that corporation tax was normally payable at an effective rate which differed from the nominal rate when the profits were £x, it seems to me reasonable to assume that the same strategies would have led to a similar result had the profits instead been £x + £y. That, in essence, is what Mr Prestwich has done, subject to the refinements of the ETR explained in his reports. I find his approach distinctly more plausible than one which assumes that the £y would automatically have increased the group's tax base by the same amount. There are a number of different ways in which the increase in turnover represented by £y might have led to an increase in the tax base of less than £y, for example if extra capital expenditure had been incurred or if budgets for particular categories of costs of sale had been relaxed. Mr Prestwich accepted in cross-examination that his approach was a generic one, based on high level assumptions; but that seems to me inevitable, given the hypothetical nature of the exercise and the paucity of the data.
447. It is fair to say that in one or two cases Mr Prestwich's approach produced some surprising looking results, including two years (1983 and 2001) where his adjusted ETR for the Littlewoods group was in excess of 50%. He was able to explain, in answer to some questions from myself, how results of this nature could occur, and in particular why the ETR for 2001 was so high. I confess to some doubts whether it can be appropriate to apply an ETR to the additional profits which is higher than the nominal rate of tax, but it is unnecessary for me to explore the point further because the result of substituting the nominal rate would be to reduce Mr Prestwich's aggregate figure for the set-off amount. The claimants are content to accept Mr Prestwich's figure as it stands, without any further adjustment in their favour.
448. To conclude, therefore, I prefer the approach and methodology of Mr Prestwich, and I accept his calculation of the set-off amount. The question is academic, however, if I am right in my view that the additional corporation tax cannot be taken into account, even as a matter of English law, when quantifying the restitutionary claims.

XII. Final conclusions

449. I have already provided summaries of many of my conclusions on the main issues in the case, as follows:

Subject Paragraphs

(1) The underlying tax issues 149

(2) Estoppel and abuse of process 252

(3) The adequate indemnity issue 302, 310

(4) Conforming construction and disapplication 341

(5) Quantum 413-417, 434, 448

450. The end result of my long and twisting journey through the interesting and difficult issues so ably presented and argued on both sides is that the claims succeed in full. In barest summary, the main reasons for this are:

(a) it is not open to HMRC to reopen the underlying tax issues, because it would be an abuse of process to permit them to do so;

(b) EU law entitles the claimants to receive an adequate indemnity for the loss occasioned to them by the overpayments of VAT;

(c) as a matter of EU law, such an indemnity requires the payment to the claimants of an amount of interest which is broadly commensurate with the loss of use value of the overpaid tax, running from the dates of payment of the tax until the dates when the loss of use value is fully restored to them;

(d) sections 78 and 80 of VATA 1994 cannot be construed conformably with EU law, and must therefore be disapplied in such a way as to permit the claimants to pursue both their Woolwich claims and their mistake-based DMG claims;

(e) as a matter of English law, the correct approach to quantification of the claims is to ascertain the objective use value of the overpaid tax, which is properly reflected in an award of compound interest;

(f) since the claimants are content to receive the use value of the overpaid tax to the Government, which is admittedly less than the loss of use value to themselves, the award of compound interest should be computed in accordance with the unchallenged rates and methodology advanced by the claimants' expert, Professor Kay;

(g) the actual benefit derived by the Government from the overpayments is irrelevant to the objective use value of the money, and even if actual benefit were the correct measure of restitution under English law, it would be precluded by EU law if the actual benefit fell short of the objective use value of the money;

(h) if, however, actual benefit were the correct measure of the restitution to be made by the Government, I would quantify it as equivalent to an award of simple interest only, following the general approach of the Government's expert, Dr Richardson; and

(i) in quantifying the claims, no account should in any event be taken of the additional corporation tax which would have been paid by the claimants if the overpayments of VAT had never been made.